Polygon 2019 by MDC-Polygon

1780 West 49th Street, Hialeah, Florida 33012, USA Editorial Note: Polygon is MDC Hialeah's Academic Journal. It is a multi-disciplinary online publication whose purpose is to display the academic work produced by faculty and staff. We, the editorial committee of Polygon, are pleased to publish the 2019 Spring Issue Polygon which is the twelfth consecutive issue of Polygon. It includes seven regular research articles. We are pleased to present work from a diverse array of fields written by faculty from across the college. The editorial committee of Polygon is thankful to the Miami Dade College President, Dr. Eduardo J. Padrón, Miami Dade College District Board of Trustees, the Hialeah Campus President, Dr. Joaquin G. Martinez, Chairperson of Hialeah Campus Liberal Arts and Sciences, Dr. Caridad Castro, Chairperson of Hialeah Campus World Languages and Communication, Professor Liliana Cobas, Director of Hialeah Campus Administrative Services, Ms. Andrea M. Forero, Director of Hialeah Campus Network & Media Services, Mr. Juan Villegas, all staff and faculty of Hialeah Campus and Miami Dade College, in general, for their continued support and cooperation for the publication of Polygon. Sincerely, Editorial Board Members of Polygon: Dr. M. Shakil (Editor-in-Chief), Dr. Jaime Bestard and Professor Victor Calderin Advisory & Reviewer Committee of Polygon: Dr. Kelly Kennedy, Dr. Alex Gancedo, Prof. Loretta Blanchette, Prof. Rene Barrientos, Dr. Melissa Lammey, Dr. Carlos Ruiz, Prof. Rodolfo Cruz, Dr. Victoria Castells, Dr. Mariana Vaillant Molina, Dr. Allison Thomas Johnson Patrons: Dr. Joaquin G. Martinez, President, Hialeah Campus Dr. Caridad Castro, Chair of Liberal Arts and Sciences Professor Liliana Cobas, Chair of World Languages and Communication Professor Charles Williams, III, Chair of Business, Engineering & Technology Mr. Alexander Hernandez, Director of Learning Resources Dr. Luis M. Rodriguez, Director of Continuing Education Ms. Andrea M. Forero, Director of Hialeah Campus Administrative Services Mr. Juan Villegas, Director of Hialeah Campus Network & Media Services Miami Dade College District Board of Trustees: Chair Bernie Navarro, Vice Chair José K. Fuentes, Trustee Dr. Anay Abraham, Trustee Michael Bileca, Trustee Marcell Felipe, Trustee Benjamin León III, Trustee Carlos A. Migoya Dr. Eduardo J. Padrón, President, Miami Dade College Dr. Lenore Rodicio, Executive Vice President and Provost, President, Miami Dade College

CONTENTS ARTICLES / AUTHOR(S)

PAGES

Participation of Women in Computer Programs at Miami Dade College - Alicia Ibarra, Rodolfo Cruz and Olga Canedo

1 - 10

Protecting First Responders in Disaster Recovery Situations with the Internet of Things: Design of a Disaster Air Quality Monitor (DAQM) Prototype - Rodolfo Cruz, Ernesto Rodriguez, Carlos Luis, Jose A. Aparicio, Emmanuel Garcia

11 - 16

Reducing Disaster Victim Search and Rescue Time with the Internet of Things: Design of a Victim Location and Assistance Rover (VLAR) Prototype - Rodolfo Cruz, Carlos Luis, Ernesto Rodriguez, Jose A. Aparicio, Emmanuel Garcia

17- 22

Controlling Rodent Infestation in the Internet of Things Era: A Design of a Smart Mousetrap Prototype - Rodolfo Cruz, Emmanuel Garcia, Jose A. Aparicio, Ernesto Rodriguez, Carlos Luis

23 - 26

Phylogenetic study of the spider community of Simpson Park in Miami, Florida through DNA barcoding

27 - 33

- Anette Sanz A Statistical-based Analysis on the Effect of the Factors Underlining the Structure of the TIMSS Instrument for Measuring Perception of Students toward Mathematics and their Teachers and Gender - Dr. Nelson De La Rosa

Statistical Analysis through Opinion Mining in the Trends of Cruise Industry - Dr. Lourdes Gonzalez and Dr. Nelson de la Rosa

Previous Editions Link: https://issuu.com/mdc-polygon

Disclaimer: The views and perspectives of the authors presented in their respective articles published in Polygon do not represent those of Miami Dade College. Mission of Miami Dade College As democracyâ&#x20AC;&#x2122;s college, Miami Dade College changes lives through accessible, high-quality teaching and learning experiences. The College embraces its responsibility to serve as an economic, cultural and civic leader for the advancement of our diverse global community.

34 - 48

49 - 53

iii Solicitation of Articles for the 2019 Issue (12th Issue) of Polygon: The editorial committee would also like to cordially invite the MDC community to submit their articles for consideration for the 2019 Issue (12th Issue) of Polygon. Guidelines for Submission POLYGON “Many Corners, Many Faces (POMM)” A premier professional refereed multi-disciplinary electronic journal of scholarly works, feature articles and papers on descriptions of Innovations at Work, higher education, and discipline related knowledge for the campus, college and service community to improve and increase information dissemination, published by MDC Hialeah Campus Liberal Arts and Sciences Department (LAS). Editorial Board: Dr. Mohammad Shakil (Mathematics) (Editor-in-Chief) Dr. Jaime Bestard (Mathematics) Prof. Victor Calderin (English) Manuscript Submission Guidelines: Welcome from the POLYGON Editorial Team: The Department of Liberal Arts and Sciences at the Miami Dade College–Hialeah Campus and the members of editorial Committee - Dr. Mohammad Shakil, Dr. Jaime Bestard, and Professor Victor Calderin – would like to welcome you and encourage your rigorous, engaging, and thoughtful submissions of scholarly works, feature articles and papers on descriptions of Innovations at Work, higher education, and discipline related knowledge for the campus, college and service community to improve and increase information dissemination. We are pleased to have the opportunity to continue the publication of the POLYGON, which will be bi-anually during the Fall & Spring terms of each academic year. We look forward to hearing from you. General articles and research manuscripts: Potential authors are invited to submit papers for the next issues of the POLYGON. All manuscripts must be submitted electronically (via e-mail) to one of the editors at mshakil@mdc.edu, or jbestard@mdc.edu, or vcalderi@mdc.edu. This system will permit the new editors to keep the submission and review process as efficient as possible. Typing: Acceptable formats for electronic submission are MSWord, and PDF. All text, including title, headings, references, quotations, figure captions, and tables, must be typed, with 1 1/2 line spacing, and one-inch margins all around. Please employ a minimum font size of 11. Please see the attached template for the preparation of the manuscripts. Length: A manuscript, including all references, tables, and figures, should not exceed 7,800 words (or at most 20 pages). Submissions grossly exceeding this limit may not be accepted for review. Authors should keep tables and figures to a minimum and include them at the end of the text. Style: For writing and editorial style, authors must follow guidelines in the Publication Manual of the American Psychological Association (5th edition, 2001). The editors request that all text pages be numbered. You may also please refer to the attached template for the preparation of the manuscripts.

iv Abstract and keywords: All general and research manuscripts must include an abstract and a few keywords. Abstracts describing the essence of the manuscript must be 150 words or less. The keywords will be used by readers to search for your article after it is published. Book reviews: POLYGON accepts unsolicited reviews of current scholarly books on topics related to research, policy, or practice in higher education, Innovations at Work, and discipline related knowledge for the campus, college and service community to improve and increase information dissemination. Book reviews may be submitted to either themed or open-topic issues of the journal. Book review essays should not exceed 1,900 words. Please include, at the beginning of the text, city, state, publisher, and the year of the book’s publication. An abstract of 150 words or less and keywords are required for book review essays. Notice to Authors of Joint Works (articles with more than one author). This journal uses a transfer of copyright agreement that requires just one author (the Corresponding Author) to sign on behalf of all authors. Please identify the Corresponding Author for your work when submitting your manuscript for review. The Corresponding Author will be responsible for the following: 

ensuring that all authors are identified on the copyright agreement, and notifying the editorial office of any changes to the authorship.  securing written permission (via email) from each co-author to sign the copyright agreement on the co-author’s behalf.  warranting and indemnifying the journal owner and publisher on behalf of all coauthors. Although such instances are very rare, you should be aware that in the event a co-author has included content in their portion of the article that infringes the copyright of another or is otherwise in violation of any other warranty listed in the agreement, you will be the sole author indemnifying the publisher and the editor of the journal against such violation. Please contact the editorial office if you have any questions or if you prefer to use a copyright agreement for all coauthors to sign. Instructions for the Preparation of Manuscripts for the Polygon: (THE TITLE IS HERE) (12 pt, bold, 32 pt above) NAME IS HERE (11 pt16 pt above, 32 pt below) ABSTRACT is here, not exceeding 160 words. It must contain main facts of the work. (11 pt) Key words and phrases (11 pt) Introduction (11 pt, bold, 24 pt above, 12 pt below) Main Body of the Article Discussion Conclusion Acknowledgements REFERENCES (11 pt, 30 pt above, 12 pt below) [1] M. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions, with Formulas, Graphs, and Mathematical Tables. Dover, New York, 1970.

v [2] J. Galambos and I. Simonelli, Products of Random Variables – Applications to Problems of Physics and to Arithmetical Functions, CRC Press, Boca Raton / Atlanta, 2005. [3] S. Momani, Non-perturbative analytical solutions of the space- and time-fractional Burgers equations. Chaos, Solitons & Fractals, 28(4) (2006), 930-937. [4] Z. Odibat, S. Momani, Application of variational iteration method to nonlinear differential equations of fractional order, Int. J. Nonlin. Sci. Numer. Simulat. 1(7) (2006), 15-27. (11 pt) Author’s Biographical Sketch (Optional): Dr. Y. Abu received his Master’s and Ph. D. Degrees in Mathematics from the University of Small Town, USA, in 1987, under the direction of Dr. M. Opor. Since 1989, he has been teaching at the Community College of Small Town, USA. His research interests lie in the Fractals, Solitons, Undergraduate Teaching of Mathematics, and Curriculum Development. (11 pt) Address: Department of Liberal Arts & Sciences (Mathematics Program), Community College of Small Town, P. O. Box 7777, Small Town, USA. E-mail: yabu@ccst. (11 pt)

Participation of Women in Computer Programs at Miami Dade College Alicia Ibarra, Rodolfo Cruz and Olga Canedo School of Engineering Technology and Design Miami Dade College â&#x20AC;&#x201C; Hialeah Campus, Hialeah Florida aibarra@mdc.edu, rcruz2@mdc.edu, olga.canedo001@mymdc.net

Abstract This article analyzes the participation of women in computer programs at Miami Dade College in the levels of Certification, Associate and Bachelor. The objective of the research is to identify the percentage of participation of women in computer programs and their academic performance. The study is relevant since the low female representation in this discipline can have a negative impact on a society that is increasingly dependent on technology. If this rate of decline continues, society will lose the presence of women in the areas of production, innovation, transfer, and development of scientific knowledge. The results obtained can later derive in some strategies that allow increasing the participation of women in the race, both at the student level and at the teaching level.

Keywords: Women in technology, MDC, Miami Dade College, Computer programs 1. Introduction. The low participation of women in technological careers has been one of the most discussed topics of education during the last years. Due to this global phenomenon and the growing need for professionals in this area, it is essential to consider: why is the career of computer science of little interest to women? and in what way can institutions address this situation? Several studies show that there is no simple answer to the fact that women do not decide to study computer careers since the reasons seem to be linked to the nature of socially defined roles. In addition, other factors such as lack of confidence among students despite their obvious skills; the lack of women teachers and role models and a culture that does not invite women to venture into computer programs (Hill, Corbett, & St. Rose, 2010). It is evident that there is a lack of women in computing, and in general lack of diversity in the human teams that work in this sector. According to a study conducted by the National Center for Women & Technology (NCWIT, 2019), in 2017, 26% represented the female workforce against 74% of men, and it is expected that by 2026, 3.5 million jobs will be opened in computing, and to the current growth only 17% of these positions can be covered. This indicates that the efforts made by various organizations have not been sufficient for a real change in this problem. As a consequence, technological development will continue to generate solutions to problems without considering all perspectives.

2. Background. In the search for reasons why women's participation in computing is so low, the Science, Technology, Engineering, and Mathematics (STEM) program was developed thinking of taking students from early stages of study to mastery and doctorate levels, and with the objective of greater the inclusion and permanence of women in this program, in addition to having a closer follow-up to identify in which stages the women defected in this study route (Soe & Yakura, 2008). The results of the STEM program show efforts oriented more towards science, mathematics, and engineering in general; and in the case of technology, it is addressed in a more general sense for college levels, that is, without going into the specific field of computation. Other results derived from the program show four relevant aspects of why women avoid becoming involved in the computer area: (1) the stereotype of computer people is that of

uncomfortable and nerdy men who lack interpersonal skills and are obsessed with technology; (2) the related image of computing as an area dominated by men oriented to work not with people but with "machines"; (3) the limited knowledge of the computer area as a discipline and as a career; and (4) the perception of topics related to computing as unattractive and / or boring (Banerjee & Santa Maria, 2013). In the context of the Miami Dade College campuses, the data of the 39 programs related to the computing area of the certification programs, and the associate and bachelor's degrees show that from 2006 to 2019, a constant registrations average has been maintained of 17% of women. This indicates that during the last 14 years the gender gap has remained constant and are compatible with the country's findings.

3. Methodology. The study was conducted with data obtained from the Institutional Research of Miami Dade College, in a period of 14 years, from the year 2006 to 2019. From the databases, only the student information of the 39 programs linked to the area was selected of computing of the certification programs and grades of associate and bachelorâ&#x20AC;&#x2122;s degree, with the most relevant attributes for this study. As a result, a data table of 5249 records was obtained, corresponding to a total of 870 records of women and 4379 records of men. This table formed the basis for the subsequent stages of the knowledge discovery process. The data table was constructed with the Excel tool of Microsoft Office. To perform the data processing, the information in the table was analyzed and a selection of various attributes was made depending on the factors that were wanted to be analyzed. Through the creation of dynamic tables in Excel (PivotTables), the selected data was reorganized and filtered to obtain the desired report and graphs. As the most relevant data for the study were selected the student's identifier, name of the campus, academic program, program description, GPA, year of admission, age range, sex, and ethnicity.

4. Analysis of Data Miami Dade College has a very flexible admission process, this contemplates that all applicants will be admitted if they meet the required level of English, while choosing the degree and career they are interested in studying. This process is carried out for three terms per year (Spring, Summer, and Fall); in which the candidate can change their career options, according to their interests and opportunities. The academic computer programs contemplate Associate and bachelorâ&#x20AC;&#x2122;s degrees and Certification programs. The student entry number varies as shown in Table 1, for various reasons: financial, student's decision or program information. Table 1. Students admitted by academic degree from 2006 - 2019

Program Female Certificate 85 Associate 708 Bachelor 77 Total 870

%F 28% 16% 19% 17%

Male 222 3821 336 4379

%M 72% 84% 81% 83%

Total 307 4529 413 5249

In this information, the percentage of women in the computer career is determined in relation to the percentage of men. These data are represented in Figure 1, which shows the proportion of female students of computing with respect to male students.

Figure 1 Students by Academic Program, 2006 - 2019 As can be seen in Figure 1, the number of women who choose the career is very low with respect to the number of men who do, is 17% on average, which is practically a ratio of 1 to 6. This indicates that the career awakens a much greater interest in men than in women, as the studies mentioned above show, and it may be necessary to make a much greater effort to promote the career to increase the motivation towards these programs of computing. Table 2 shows the number of students belonging to the computer career classified by gender from 2006 to 2019. Table 2. Students admitted by gender from 2006 â&#x20AC;&#x201C; 2019

Year Unknown 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Total

Female 260 4 7 9 15 10 21 15 25 34 58 111 120 158 23 870

%F 30% 0% 1% 1% 2% 1% 2% 2% 3% 4% 7% 13% 14% 18% 3% 17%

Male 1088 15 46 35 40 56 70 91 139 203 300 505 676 969 146 4379

%M 25% 0% 1% 1% 1% 1% 2% 2% 3% 5% 7% 12% 15% 22% 3% 83%

Total 1348 19 53 44 55 66 91 106 164 237 358 616 796 1127 169 5249

The data in Table 2 show a disproportion between the number of women and men entering the Computing career. It can be seen that the proportion of women's registration with respect to men is on average 17%, and supports the difference in female representation in relation to men. It is important to note that the behavior of the number of women in the career through the years has an average sustained trend; that is, it does not show a downward trend. As of 2012, there is a sustained upward trend (Figure 2), possibly associated with the increase in the number of students admitted.

Figure 2. Students of the Female Gender, 2006 to 2019 When thinking about computing, one of the main skills required for any student is logical thinking, which is considered essential for success in this career. Historically, men have surpassed women in this area, but in recent years the gender gap has been reduced, and today women are doing well or even better than men. The grades for women are being slightly higher as we can see in Figure 3 for Miami Dade College students.

Grade Point Average in Computer Program, by Gender, 2006â&#x20AC;&#x201C;2019 GRADE POINT AVERAGE

3.5 3.0 2.5 2.0 1.5

Female

1.0

Male

0.5 0.0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 SCHOOL ENTRY YEAR

Figure 3 Grade Point Average in Computer Program, by Gender, 2006â&#x20AC;&#x201C;2019

It should also be considered that the distribution of students integrated into the 39 computer programs at the certificate, associate and bachelor levels offered by the MDC has a higher population in the Associate degree (Figure 4). Where the greatest coverage is with the Computer Science, Computer Arts Animation, and Computer Information Systems programs from 2006 to 2019.

Figure 4 Students by Associate In the bachelorâ&#x20AC;&#x2122;s degree, the highest percentage of students is in the BS Information Systems Technology program (Figure 5). For certification programs, the largest number of students is in the Computer Aided Design Operator for CRTFA, and Computer Aided Design Assist for CRTNF (Figure 6).

Students by Bachelor

22 7

40 18

160 140 120 100 80 60 40 20 0

Male

139

Female

Figure 5 Students by Bachelor

Students by CRTFA

3 0

5 0

8 3

14 3

70 60 50 40 30 20 10 0

Male

Female

20 12 10 9

35 30 25 20 15 10 5 0

Male

Female

Students by CRTNF

Figure 6 Students by CRTFA and CRTNF In the different academic programs and degrees, the difference of integrated women in relation to men is widely marked. Also, without considering the differences by gender, it is clear that there are programs that must be reviewed and evaluated to determine if they are kept as part of the educational offer offered by the MDC, or if these are already included in the Computer Science program. Comparing these results with the predictions of job growth in the computing field (Figure 7) analyzed by the Bureau of Labor Statistics in 2018, where Software Developers are expected to be one of the ten fastest growing occupations in the world next eight years. (DuBow & Pruitt, 2018)

Predicted Job Growth by Computing Occupational Classification, 2016-2026 35% 30% 25%

31%

29%

20% 15% 10% 5%

19% 15% 12%

11%

0% -5%

-7%

-10%

Figure 7 Predicted Job Growth of Select Computing Occupational Classifications (2016-2026) The Computer Science program of MDC's Associate of Arts aims to teach students how to design technical programs, research, create new technologies, develop operating systems, code device drivers, write specialized programming languages and implement complex applications in a variety of configurations (Miami Dade College, 2013). Therefore, MDC could be working to generate students in the field of higher job growth that is the developer of software applications. However, according to the Occupational Outlook Handbook, a software developer addresses the need for new applications in smartphones and tablets, as well as more specific industries such as health. Software developers will see new opportunities due to an increase in the number of products that use software (Bureau of Labor Statistics, 2019). For example, more computer systems are being incorporated into consumer electronics and other products, such as cell phones and home appliances. All this leads to new concerns about threats to computer security, which could result in increased investment in security software to protect networks and electronic infrastructure. It is considered then, that the increase in the consumption of customized software, will increase the demand of software developers. As a result, computer science would have to consider the necessary skills related to the industry in greatest demand in their academic programs, and students must consider as important classes all those related to software development, learning how technological changes and programming languages occur. Students from historically disadvantaged groups, such as Hispanics or African Americans, both men and women have less opportunity to access advanced computer courses, which would negatively affect when deciding what career to choose the University and thus successfully complete a career in computing. However, the participation of Latinos in this field is widely represented by both men and women, followed by non-Hispanic blacks as shown in Figure 8. It is important to mention that the number of Hispanic students of the female gender exceeds in amount if compared to the other ethnic groups. It should be kept in mind that the large representation of Hispanics in the computer program is not indicative of greater success in the workplace.

Students by Racial -Ethnic Groups, by Gender, 2006 -2019 4000 3334

3500 3000 2500 2000

Female

1500

Male

1000

602

500 2

124

648 7

103

245

Figure 8 Students by Racial-Ethnic Groups, by Gender, 2006-2019

5. Discussion. The data show that the reasons for a career choice of students are probably outside the school context, and the marked difference between men and women in the area of computing is defined from the early stages of education. In this sense, actions that motivate high school students should be reinforced in the choice of a career in the computer science area. If this situation is not changed, the discipline may lose the contribution that sex feminine provides in terms of skills and different perspectives. However, a valuable result is that at Miami Dade College there has not been a downward trend in the entry behavior of women into the career in recent years. In fact, starting in 2012, there has been a sustained tendency to rise. In this context, it is important to determine actions that allow the phenomenon to be maintained. The results also conclude that the number of women in the computer field would lead us to think that men are superior to women and adapt better to this field of work, but it is shown that women are the same or even higher in academic grades according to the comparative data of the GPA of 2.8 for women against 2.5 for men. Regarding the first question posed at the beginning of the article: What is the representation of women in the computer programs of the MDC? We can conclude that the number of women in the race in relation to the number of men is in the average order of 17% women -83% men, which confirms the low female representation in the career, as documented in the Studies conducted by the country. Regarding the second question: How can institutions address this situation? There are several alternatives to be considered to reduce this wide difference between women and men in this career. ď&#x201A;ˇ

ď&#x201A;ˇ

Considering that engineers in computer systems are made in colleges and universities, it is important to make small changes such as providing a broader view of the 39 computer programs in the introductory courses. Implement special work schemes that include more female professors to promote the integration of more female students.



Implement induction courses that showcase of success of women of diverse ethnic groups in the computer career and the route of success through which one must travel to reach the occupations demanded by the labor market. Having information is the power to make better decisions. Design favorable working environments, with technological advances it is possible to visualize the development of remote work from home, favoring work schedules with your personal life.

6. Future Work. The study carried out represents an analysis of the situation of women in the career at the MDC. More comprehensive and complex studies are needed to analyze some other more qualitative elements, such as previous experience in programming, the results of the programming courses, the concept of self-efficacy that women have and the possible impact of feminine or masculine teaching roles. Additionally, more work should be done to establish strategies that allow women a greater entrance to the career, as well as a greater permanence in it. Also, a better understanding of the phenomenon in the programming area will allow establishing pedagogical strategies that contribute not only to help the permanence of women but also to increase the success and reduce the desertion of students of both genders.

7. Conclusion. Women have made impressive advances in computing, but they are still a minority in relation to men. Derived from the analysis of information, it is suggested the creation of environments that support the achievements and interest of the female gender in computer programs, which will encourage them to continue in this career. Understanding this situation better provides important information for decision making, not only to improve the conditions of women in the career but also to understand that some decisions could be made to increase the number of students in the area. This problem is of institutional, national and international interest.

Acknowledgments The authors are thankful to the editorial committee and reviewers for their comments and suggestions, which improved the quality of the paper.

8. References [1.] Ashcraft, C., Eger, E., & Friend, M. (2012). National Center for Women & Technology (NCWIT). Retrieved April 10, 2019, from www.ncgs.org. [2.] Banerjee, S., & Santa Maria, R. (2013). A study of students’ perception of computer education: Lack of interest in STEM fields for female students. International Journal of Technology, Knowledge and Society, 8, 93–106. [3.] Bureau of Labor Statistics. (2019, April 12). Retrieved from Occupational Outlook Handbook: https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm [4.] DuBow, W., & Pruitt, A. (2018). NCWIT Scorecard: The Status of Women in Technology. Boulder, CO: NCWIT. [5.] Hill, C., Corbett, C., & St. Rose, A. (2010, February). www.aauw.org. Retrieved April 21, 2019, from www.aauw.org.

[6.] Miami Dade College. (2013). Retrieved http://www.mdc.edu/computerscience/#collapse2

from

Academics:

Computer

Science:

[7.] NCWIT. (2019, March 3). National Center for Women & Technology. Retrieved April 20, 2019, from https://www.ncwit.org [8.] Soe, L., & Yakura, E. K. (2008). Whatâ&#x20AC;&#x2122;s wrong with the pipeline? Assumptions about gender and culture in IT work. Women's Studies, 37, 176â&#x20AC;&#x201C;201.

Protecting First Responders in Disaster Recovery Situations with the Internet of Things: Design of a Disaster Air Quality Monitor (DAQM) Prototype Rodolfo Cruz, Ernesto Rodriguez, Carlos Luis, Jose A. Aparicio, Emmanuel Garcia School of Engineering Technology and Design Miami Dade College – Hialeah Campus, Hialeah Florida rcruz2@mdc.edu, ernesto.rodriguez035@mymdc.net, carlos.luis003@mymdc.net, jose.aparicio002@mymdc.net, enmanuel.garcia002@mymdc.net

Abstract After natural disasters, there is a great risk of gas emissions and fires that can be fatal, not only for the victims of the disaster and the residents in the affected area, but also for the rescuers. DAQM is a device that provides rescue personnel with real time information about the safety of a disaster scene to prevent unnecessary deaths due to fires, fuel and gas leaks caused by the disaster. This tool analyzes the environmental conditions and transmit telemetry to help rescuers make decisions on how to proceed before any human being is put in harm’s way.

Keywords: disaster, rescue, sensors, hexapod, robot, DAQM, servo, motors, Arduino, ESP8266, MQ-3, MQ-5, ultrasonic, photoresistors, GL55

1. Introduction Natural disasters are one of the biggest problems mankind has faced since the beginning of history. They have a big impact in the economy and people’s lives. The frequency of geophysical disasters (earthquakes, tsunamis, volcanic eruptions and mass movements) remained broadly constant throughout the period between (1994 – 2013) (Centre for Research on the Epidemiology of Disasters (CRED)). After natural disasters, there is a great risk of gas emission and fires that can be fatal not only for the victims of the disaster and the residents in the affected area but also for the rescuers. This project “Disaster Air Quality Monitor” (DAQM) is oriented to provide a way to avoid unnecessary death in fires and natural disasters. In order to prevent the risk of losing human or animal lives, DAQM can enter the affected sectors and communicate to the rescue team the environmental conditions of the place (humidity, temperature, air composition, liquid fuel, etc.) in real time. Thus, providing the rescue personnel with valuable information they will need before to prepare for the incursion.

2. Background and Motivation As students in charge of making a functional project, we came up with the idea that we could focus our work on helping people and make easier certain tasks that, at this time, were very difficult for us to do without putting ourselves at risk. One of the problems that we tried to solve was reducing the number of

lives that were lost in each natural disaster due to the destruction they left in their path and the difficulty of accessing certain areas in some without putting others at risk. DAQM was created with that goal in mind and was focused in the field of toxic gases and liquid fuels detection. After an arduous investigation it was decided that this would be its main function, that way it would serve to warn the rescue units in what conditions the site was explored. Everything we hoped to achieve with this project was to help people in extreme situations. Something that could warn us of the dangers and how to avoid them, so that we or the animals are the ones who are exposed to areas where we should not be due to the levels of toxicity in the environment. This was our main motivation throughout the development, helping people.

3. Technical Description DAQM was built using IoT technology. this technology involves a wide variety of sensors, motors, and boards; which can communicate with each other making it easier to fully operate. For the body of the robot, a pre-constructed prototype of a spider was used, to which different sensors were added, in order to reach its current stage of development. Next, each sensor, board, and part used in DAQM will be broken down by sections. 3.1. Hexapod Robot: The body of DAQM was obtained through the company "FreeNove", which offered a product that was close enough to what we were looking to start.

The body was a mountable kit, which came with its instructions to begin its proper initial operation.

Figure 4: Arduino Mega with Some Modification (The Board That We Are Using In DAQM)

3.2. Servo Motors

Figure 1: Hexapod Robot Kit

A servo-motor is an actuator with a built-in feedback mechanism that responds to a control signal by moving to and holding a position, or by moving at a continuous speed.

Figure 5: Servo Motors used on DAQM

Figure 2: Hexapod After Assembly

The Arduino Uno is a board for microcontrollers based on the ATmega328. It has 14 digital input / output pins (of which 6 can be used as PWM outputs), 6 analog inputs, a 16 MHz ceramic resonator, a USB connection, a power connector, an ICSP header and a reset button. It contains everything necessary to support the microcontroller; simply connect it to a computer with a USB cable or turn it on with an AC to DC adapter or a battery to start. The One differs from all previous cards in that it does not use the USB to FTDI serial driver chip. Instead, it presents the Atmega16U2 (Atmega8U2 up to version R2) programmed as a USB to serial converter.

Figure 3: Different types of Arduino Boards

Figure 6: Pinout of a Servo Motor

3.3. ESP8266 Wi-Fi Module: ESP8266 Wi-Fi Module offers a complete and autonomous Wi-Fi network solution; It can be used to host the application or to download Wi-Fi network functions from another application processor. ESP8266 Wi-Fi Module is one of the most integrated Wi-Fi chips in the industry; integrates antenna switches, balun RF, power amplifier, low noise reception amplifier, filters, power management modules, requires a minimum of external circuits, and the entire solution, including the front module, is designed occupy a minimum area of PCB.

Figure 7: ESP8266 Wi-Fi Module

Figure 8: ESP8266 Pinout

3.4. Ultrasonic Module HC-SR04: Ultrasonic ranging module HC - SR04 provides 2cm -

Figure 12: MQ-3 Sensor

400cm non-contact measurement function, the ranging accuracy can reach to 3mm. The modules include ultrasonic transmitters, receiver and control circuit.

Figure 13: Structure and Configuration

Figure 9: Ultrasonic Module HC-SR04

3.6. MQ-5 Sensor: This is a very simple sensor to use to detect LPG, natural gas, coal gas. Ideal for measuring concentrations of LPG (composed mostly of propane and butane) in the air.

Figure 10: Ultrasonic Transmitter/Receiver Figure 11: Ultrasonic Module Pinout

3.5. MQ-3 Sensor: The MQ-3 alcohol sensor is suitable for the detection of alcohol concentration, just like a common breathalyzer. It has a high sensitivity and fast response time. This provides an analog resistive output based on the concentration of alcohol. Avoid the noise of kitchen fumes, alcohol and cigarette smoke. The sensitivity can be adjusted by the potentiometer. The sensitive material of the gas sensor MQ-5 is SnO2, which with lower conductivity in clean air. When the target fuel gas exists, the conductivity of the sensors is higher along with the concentration of gas that rises.

Figure 14: MQ-5 Sensor

3.7. Photoresistors GL55: Photoresistor is a resistor which made of semiconductor material, and the conductance changes with luminance variation. The photoresistor can be

Figure MQ-5 Configuration Figure 17:15: Specifications Figure 16: GL55

manufactured with different figures and illuminated area based on this characteristic.

4. Project Timeline DAQM is a project which was carried out following several phases which helped us to organize everything in order to focus it and direct it towards the desired direction. The first phase of our project was generally based on the search of data and information that will help us to decide what sensors we would put to the robot, because our robot should be focused on a specific problem in order to make it more efficient when working with them. Some of the data obtained about the deaths produced by natural disasters showed that a large part of the rescue team members died or got sick from entering areas where there was a great concentration of toxic gases, so we decided to guide

Table 1: Global annual death rate from natural disasters, by decade (1900s-2010s)

our project for that. way to avoid this problem.

Table 2: Fire deaths 2008-2017 (U.S. Fire Administration (USFA))

Table 3: Negative Effects of a Fire 2017 (U.S. Fire Administration (USFA))

After collecting all the necessary data and deciding what direction our project was going to take, we started the second phase. During this period, we focused on finding a robot that could be useful for our purpose, we needed something small, with enough space to place several people, and able to move in places of difficult access. For that reason, we chose the "Hexapod Robot Kit" by FreeNove. This robot will allow us to make all the modifications that we needed and could move in difficult places. Before starting to put the sensors and modifications; we are dedicated to assembling it and modify all its base code to do what we were looking for in specific. This work took us several weeks because different tests were done to calibrate the force applied by the motors and the speed of the movement. Once we had perfected the code of the robot, we began the search of what would be the sensors that we would use in the project so that it would fulfill the objective. After several days comparing existing sensors in the market, we decided to buy sensors that could detect gases, liquid fuels, lighting sensors (for a future solar cell that helps to charge batteries) and the ultrasonic sensor for the evasion of obstacles that we It was going to be very useful to prevent it from getting stuck in places with no way out. At the beginning of the last phase we had everything planned on how we were going to do things, you had the goal of the project, the robot worked perfectly, and we knew how each sensor worked. What we were missing was to integrate them into the robot's body and add the code of each sensor to the code that we had made of the spider's movement. After a couple of months working on it, we managed to have everything working perfectly, each sensor was receiving the information from the environment, the ultrasonic sensor worked in harmony with the motors of the legs, thus avoiding shocks with obstacles. We had managed to build a functional robot that would warn us of the conditions of the air at each moment. This had been our first fully functional model.

readings obtained, in order to calibrate everything and make them work in the most optimal way. Another of the sensors that had to be tested several times was the ultrasonic sensor because in the early stages of development it failed a lot and hit the obstacles, as we progressed we improved this defect until we managed to avoid every obstacle that was presented to it. front with 100% efficiency.

6. Discussion and Future Work

Figure 18: DAQM

Figure 19: DAQM

5. Testing DAQM was a project which had to go through several stages of testing, to be able to calibrate all the sensors and engines that use DAQM. To test each sensor, we built a remote control that communicated with the robot by radio frequency. Thanks to this control we were able to test their movement more precisely which made it easier to calibrate each failure that each motor could have individually, because each one needed to move very precisely in order to generate movement. The sensors used by the robot are sensors that capture readings of the surrounding environment, if they were to be used to prevent people from entering toxic or dangerous environments, each sensor must be working perfectly. The temperature and humidity sensor were subjected to different situations to see if the readings it gave were accurate. In order to test the gas and liquid fuel sensors, we bought specific sprays that would make several sensor readings. All these tests helped us to understand the sensors and the

Because DAQM is a prototype even this subject to different variations that may arise over time. The next step in the development of DAQM will be to take it to the air, this idea came to mind because the drones are much faster than any vehicle that moves on the ground and can also access different areas which adds much diversity to the project. The plan is to learn well all that is needed to make a drone work properly and to modify its code from scratch so that it can do what we need just like we did with the hexapod robot. The process to which the drone is going to be submitted will be very similar to the one that was initially done because the only thing that has been thought of in several is the movement of DAQM. The biggest challenge will be to understand the drone in its entirety and make it autonomous. For the moment it is the only thing we are working on now for a 2.0 version of DAQM. After completing this version, we hope that it can be used in real or simulated conditions to know its efficiency in these cases.

7. Conclusion To conclude, this project was a unique experience in life, because it was the first time that we all worked on a project as a team. DAQM was a project that helped us all grow as we not only worked with software but also with hardware which for many of us was totally unknown the work with hardware. We hope that DAQM can be a project that helps rescue teams in natural disasters in the future. It would be a dream to see that our small group project becomes something that we see as a tool that rescue teams usually use and help reduce the number of victims in these cases to the minimum possible.

Acknowledgments The authors are thankful to the editorial committee and reviewers for their comments and suggestions, which improved the quality of the paper.

8. References [1] Centre for Research on the Epidemiology of Disasters (CRED). "The human cost of

16 natural disasters perspective." 2015.

2015:

global

[2] U.S. Fire Administration (USFA). U.S. fire statistics. Emmitsburg, MD: U.S. Fire Administration (USFA), 2019.

Reducing Disaster Victim Search and Rescue Time with the Internet of Things: Design of a Victim Location and Assistance Rover (VLAR) Prototype Rodolfo Cruz, Carlos Luis, Ernesto Rodriguez, Jose A. Aparicio, Emmanuel Garcia School of Engineering Technology and Design Miami Dade College – Hialeah Campus, Hialeah, Florida rcruz2@mdc.edu, carlos.luis003@mymdc.net, ernesto.rodriguez035@mymdc.net, jose.aparicio002@mymdc.net, enmanuel.garcia002@mymdc.net

Abstract VLAR is a robot enhanced with Interne of Things (IoT) capabilities. It is designed to aid in finding victims trapped under debris, rubble or damaged structures left behind by natural disasters, such as earthquakes or hurricanes. VLAR’s capabilities could help first responders save time locating trapped victims, especially those still alive. It will also help to avoid putting said first responders in harm’s way when their presence is not needed in a particular location. This prototype includes IoT technologies such as communication, sensors, actuators, dc motors, Arduino and Raspberry Pi which are part of a new CCC in IoT Applications program at Miami Dade College – Hialeah Campus. VLAR is in early stages of development and still needs to be improved and tested in real disaster conditions.

Keywords: disaster, debris, robot, IoT, Internet of Things, MDC, victim, CCC, Miami Dade College 1. Introduction Natural disasters have a big impact in the economy and people’s lives. The frequency of geographical disasters (earthquakes, tsunamis, volcanic eruptions and mass movements) remained broadly constant throughout the period between (1994-2013). Death rates, on the other hand, increased over the same period (Centre for Research on the Epidemiology of Disasters (CRED)).

Figure 2: Global Reported Natural Disasters by Type (Ritchie and Roser)

Figure 1: Number of Recorded Natural Disasters, All Natural Disasters (Ritchie and Roser)

Additionally, every year there are incidents in which people lose their lives during rescue and disaster relief activities. The VLAR project is focused towards helping to minimize the loss of life caused by these disasters by using robotics to scout areas in which direct access by first responders might not be feasible, such as under debris, rubble, or damaged buildings. It is precisely in these places where victims end up trapped in places that makes them hard to locate. Professionals such as firefighters, policemen and paramedics could benefit from VLAR’s capabilities

by avoiding being put in situations where it could be dangerous to proceed into an area where the conditions of the scene are either unknown or will be putting the rescuer in danger as well as the victims that might be trapped inside.

2. Background and Motivation In the event of a hurricane, earthquake, tsunami, or other natural disaster that could leave people buried under destruction, first responders refer to the window of time when a victim has the greatest chance of survival as the “golden hour.” (Quinn). It is critical that victims be located during this precious time to increase their chances of survival. With this in mind, we came up with the idea of creating VLAR to speed the time it takes to locate victims in hard to reach areas under debris or in highly damaged structures. In addition, VLAR can keep first responders out of harm’s way.

Figure 3: Frame 1

3. Technical Description The VLAR project has built using IoT capabilities in order to amplify the basic functionality of a normal autonomous car into something that would help in more serious scenarios. VLAR started out as a basic kit composed by just the frame, wheels, motors and the main board to control the software that would run the system. Furthermore, we added more sensors, modules and separate systems to address the issues we thought rescuers would encounter somewhat often. This section will contain the labeling, details and functionality embedded in some of the key component. The frames (shown below) were used to protect and assemble the components on a sturdy surface which will increase the stability of some of the sensors. They are made of light yet firm plastic.

Figure 4: Frame 2

The Arduino board is responsible for controlling and executing most of the mechanical movements of the system. It controls the motors, servo, and the infrared sensor (used to manipulate the rover manually). This component is the one of the main powerhouses of the project.

Figure 5: Arduino Uno board

The DC motors are the main mechanical module in charge of moving the rover. They can achieve considerable speeds and the rover will have no trouble at the time of stopping. The motion will be controlled through software using the other mentioned modules.

Figure 6: DC motor

Raspberry Pis are full-fledged computers with less power than a common desktop one, yet powerful. This model specifically was the smallest one we could find to make the car as compact as we could. To begin with, we used it to control a high quality 3x3x2 camera.

Figure 8: Raspberry Pi Camera

The ultrasonic proximity sensor module uses sound waves to calculate the distance between itself and an object right in front of it. For our purpose we just needed it to detect obstacles at 15-20 centimeters. If the proximity threshold was triggered by an obstacle, then VLAR stops and, with the help of a servo motor, we rotates the proximity sensor and check which direction has less obstacles in its way to continue moving.

Figure 7 Raspberry Pi Zero W board

This camera was carefully chosen to maintain VLAR as compact as we could. It has a 1080p lens. This camera allows us to scout the terrain without the necessity of having visual contact with VLAR. To control this camera, we are using the Raspberry Pi Zero W running a Python-3 script (programming language) to buffer the video output into a local port on the network; this allows us to have the same videostream on every device connected to the same network as the Raspberry Pi.

Figure9: HC-SR04 Ultrasonic Proximity Sensor

The servo motor (shown in the picture) has the simple functionality of rotating the Ultrasonic Proximity Sensor in order to avoid obstacles in the way of VLAR. At some point of the development of VLAR it was also used to rotate the camera to have more control over the terrain.

Figure 10 Servo Motor

The IR (infrared receiver module) receives an infrared signal (such as the one used by regular TV remote controls) allowing us to remotely control VLAR. We built a controller and programmed the different buttons to activate and control different functionalities of VLAR.

It needed a long operational range, so rescuers could be far from danger.

We needed to implement enough features that would help the rescue and work effectively.

In the beginning, we bought a kit that came with the basic sensors and main frames. After assembling it all together and testing it out, we realized that the reference code needed to be tweaked a lot to fit our necessities. The code was written from scratch to make the responsiveness a lot sharper and add implementation to new features. Using the IR receiver, we designed a controller for the motors, thus manipulating the movement of VLAR. Once we had everything working properly, more planning on upcoming features was done.

Figure 11 IR Receiver

4. Project Timeline

Figure 12: Remote Control

VLAR started out without a name or idea besides with the purpose of helping society in some kind of way. We looked at natural disastersâ&#x20AC;&#x2122; statistics to find out what was a feasible for our current knowledge, skillsets and budget. After looking at natural disastersâ&#x20AC;&#x2122; death rates and the increase in occurrences as time passed, we decided to create something that would help to minimize loss of life and number of injured people. After several meetings discussing who our project would be targeted to, it concluded that VLAR would stand for Victim Location and Assistance Rover; a prototype device to help locate victims after disasters.

Figure 13: Joystick

VLAR went through several stages but, our final requirement achievements we set were: 1.

It had to be as compact as possible.

Having a camera was essential in a rescue kit, so we used the smallest Raspberry Pi model we found to go along with a small camera with night and day vision. In the early versions of the development a lot of time was spent trying to minimize the delay of the video

output, so it would be possible to maneuver the rover through the camera. Finally, we decided to establish the video output on a networkâ&#x20AC;&#x2122;s local port, that way multiple people could be watching it on different devices at the same time with less than 300 milliseconds of latency in real time. At some point during the development of VLAR, an ABS (Acrylonitrile Butadiene Styrene) arm was built using a 3D printer to help the rover with obstacles or grabbing items or evidence. Since this arm was using many servo motors, it drew a lot of power and more batteries were needed to maintain the rover fully powered for a long period of time. The arm was also quite big, therefore, added more weight and size to the rover. These two points were against our initial goal which sought a compact and efficient system and it was decided to remove the arm and later implement something more effective.

Figure 16: VLAR - Top Front View

In our late-stage of development we thought of how to make VLAR smaller. The camera was moved to the front to occupy less space on the top frame as well as to give a better perspective.

Figure 17: VLAR - Front View

5. Testing

Figure 14: VLAR - Side View

We wanted to ensure consistency with all the sensors, modules and signals. A lot of testing was done to find the weaknesses of our sensors or what could be improved in the code. A lot of changes were made such as the reduced delay on the proximity sensor to make the reaction of the rover more responsive whenever it detected an object in the front and when calculating which route to go after one was found. Additionally, testing the range capability of the infrared receiver along with the camera given a mobile hotspot as network was another important task as it would be catastrophic to lose control in the middle of a rescue operation. We made sure that the delay between these functions was as little as possible to make a better experience.

Figure 15: VLAR - Top Side View

6. Discussion and Future Work As VLAR was only a prototype, we are glad that IoT can be used not only to improve the comfort of daily tasks that could be automated, but also to save lives

and contribute to society. As we work on VLAR in the future, there are a lot of features we would like to add such as improving the working conditions in rougher surfaces. Aside from VLAR, our team was very excited throughout the time we had to work on this and are glad to have received more project ideas and support from our Miami Dade College campus to continue working on projects like these.

7. Conclusion We all felt privileged to be the first students of this IoT certification course as we will all soon be graduating, and this project was of great experience, not only working with IoT technologies but also working in a team, which required a lot of responsibility. This project greatly helped us develop our skills using IoT technologies and also expand to others to make everything we needed work properly. In the future we would like to work on more projects that help trivial issues in society or specific cases of people in need that require special automatization.

Acknowledgments The authors are thankful to the editorial committee and reviewers for their comments and suggestions, which improved the quality of the paper.

8. References [1] Centre for Research on the Epidemiology of Disasters (CRED). "The human cost of natural disasters 2015: a global perspective." 2015. [2] Quinn, Kristin. "Heartbeats in the Rubble." Trajectory 3 May 2017. Web Article. 14 May 2019. [3] Ritchie, Hannah and Max Roser. Natural Disasters. 2019. https://ourworldindata.org/naturaldisasters. 14 May 2019.

Controlling Rodent Infestation in the Internet of Things Era: A Design of a Smart Mousetrap Prototype Rodolfo Cruz, Emmanuel Garcia, Jose A. Aparicio, Ernesto Rodriguez, Carlos Luis School of Engineering Technology and Design Miami Dade College – Hialeah Campus, Hialeah Florida rcruz2@mdc.edu, enmanuel.garcia002@mymdc.net, jose.aparicio002@mymdc.net, ernesto.rodriguez035@mymdc.net, carlos.luis003@mymdc.net

Abstract Mice and rodent infestation are not only an inconvenient nuisance, it also comes with significant health risks and challenges. Pest control is an important part of our lives that will eventually be changed by the tides of the Internet-Of-Things revolution. The Smart Mousetrap prototype discussed in this paper shows how network connectivity can create smarter pest control methods for both domestic and corporate purposes, while also opening a window into the future of pest control.

Keywords: pests, mice, mousetrap, internet-of-things, IoT, mouse, control 1. Introduction The Smart Mousetrap is an experimental IoT (Internet of Things) pest control device that enhances the systems generally found in classical mousetraps but utilizes IoT capabilities and the processing power of a micro-computer to expand its features. Our first objective was to create a mouse-trap that could use the internet to send a phone notification to the user about its status. The advantage of this over a regular mousetrap would shine best in situations when the trap is deployed in a difficult-to-reach area that would require an inconvenient amount of time to reach in order to check. This would make the system advantageous to use in large areas such as warehouses and supermarkets, where mice infestation tends to be a much bigger concern than in domestic use. Initially our system simply trapped a single mouse with a single trapdoor that was controlled by an infrared sensor, but eventually we realized that much more could be done with the system. We expanded its features from a trap with a single door that could trap one mouse at a time, to a system of cells with several doors that could trap up to four mice and notify the user of cell availability.

United States could reach 10 billion (“Pest Control Market to Reach $10B”, 2016). This large number does not come as a surprise when we consider the incredible ability of pests, such as mice, to grow exponentially. One female can bring about 50 pups in one year depending on whether conditions are favorable or not (“How Fast Do Mice Multiply”, 2018) .Moreover, mice infestation also presents a serious health concern such as the spread of dangerous diseases such as leptospirosis1 and the ratbite fever (Bodkin, 2018). Mousetraps are a piece of human ingenuity that have been instrumental in the battle to stop the growth and spread of these little villains. With the IoT revolution in the horizon, it is only a matter of time before mousetraps as we know them are changed forever. Our purpose with this project is to explore what is possible to do with a mousetrap that is not only operated by a inexpensive and compact microcontroller (for context, the controller is smaller than the trap itself), but also connected to a network and able to send information about its status without the need to physically check it.

3. Technical Description

2. Background and Motivation Initially, our purpose was to study how we could enhance an everyday life object with the power of IoT. Having worked in my family’s warehouse for many years, we knew just how much of a nuisance pest like mice could be. An extraordinary amount of money is spent in pest control, and it is expected to grow exponentially. It is estimated that by 2020, structural pest control in the

Leptospirosis is a bacterial disease that affects humans and animals

Our initial project was relatively simple, the objective was to build a trap with two main capabilities: to trap mice and to use network connectivity to send a phone notification at the time the mice is caught. We were able to achieve this by linking together an infrared sensor and a servo motor attached to a trap door all inside a wooden maze that would trap mice inside the moment the infrared sensor was activated. The bridge between connecting all our devices was an Arduino microcontroller, which processed the information fed to it by the infrared sensor, and activated the servo attached to the trapdoor once the sensor was activated. Once we had this, we then used Blynk2, an internet-of-things platform, to establish a link between our trap and the phone. This allowed us to receive real-time phone notifications of the status of the trap, and whether a mouse had triggered it.

the door system in Figure 1. It consists of a 3D printed door that is glued to a hole that leads into the trap. The door rests on the servo arm and once the infrared sensor detects a mouse, it commands the servo arm to rotate its angle of repose so that the door loses its support and goes into free-fall.

Having built a trap that not only detected and caught mice, but also delivered real time information about its status, we accomplished our original objective and then we set out to improve it. The first question we asked ourselves: given what we had, how could we take it one step further? After going back to the drawing board for some time, we came up with an idea that truly made our project unique and stand out over similar projects.

Upon resetting the system, the motorâ&#x20AC;&#x2122;s angle returns to its first angle, which permits the user of the trap to set the door to rest on it once again. Then it stays in this position until the infra-red sensor inside the trap detects a mouse. The speed of free-fall is quick enough and does not give the mouse any time to react before it is trapped inside. This system proved to be reliable enough for us to implement it as our first version.

The idea was simple: we could enhance the program with a computer algorithm that would allow it to trap more than one mouse at a time. To do this, we would first need to rethink how we approached the trap-door system. No longer we could rely on a one-time activate trap, but one that could open and close the main door as needed. Then we would add several cells, each with individual trap-doors coordinated by a central algorithm that would ensure that the trap could catch at least four mice before reactivating it.

Finally, once the door was functioning we used the Blynk library to make it deliver phone notifications to the phone through the Blynk app. This allowed us to keep track of the mouse-trap in real time and enabled us to determine with accuracy at which time the mouse had been caught. More about this is talked about in the testing section.

Figure 1: Diagram of door functionality in phase one.

4. Project Timeline Work in our project was divided into three major phases. We implemented, tested it, and once we saw it worked, we went back to the drawing board to expand into what we have built. The result of this process of iteration is the current version, which allows us to trap up to four mice before we have to restart the trap.

4.1. Phase One The challenge of phase one was figuring out a door system that would react quickly enough so that once it began to close the mouse wouldnâ&#x20AC;&#x2122;t have time to react and escape. The first experimental system was a motor that pulled a string once the infra-red sensor detected something, but that system proved too slow to effectively catch mice. Down the line, we created 2

https://blynk.io/

4.2. Phase Two Figure 2 The design for the second mousetrap.

Phase two finally delivered a project that was more reliable and easier to use. Instead of using infrared

sensors that detected the heat of living creatures (in this case, mice), we replaced the system of detection with pairs of laser sender/receiver sensors that sent signals to the Arduino when the laser beam was interrupted. Instead of relying on doors that would fall due to gravity, we built a system that would have the doors attached to the servo arm, which rotated upon receiving the signal from the sensor attached to it. We also added an algorithm that would serve the purpose of guiding the mice one by one to each of the four cells in the trap, ensuring that all four of them were occupied by a mouse. This would put our system ahead of the curve by being able to trap multiple pests at a time. The only challenge we ran into was that the Arduino didnâ&#x20AC;&#x2122;t have enough power to control three servo arms, three sensors, and three lasers. This required an auxiliary power source to be connected to power up the lasers.

Once the mouse is securely trapped in cell two, the main door (M) to cell four opens again.

Mice number three enters cell four, the sensor detects it and the main door (M) closes.

Eventually the mouse walks to cell number three, the sensor detects it and then door (C) closes, trapping it inside.

10. Once the third mouse is securely trapped in cell three, the main door (M) opens again. 11. Finally, the fourth mouse enters through the main door (M), the sensor detects it, and the main door is permanently closed until the owner of the trap takes out the mice that had been caught.

The laser receiver sensor constantly sends a high reading to the digital pin in the Arduino, so long as it is receiving a direct beam from the laser emitter. However, it is important that the laser emitter is receiving enough power otherwise the beam is not bright enough to trigger a high reading from the sensor. We discovered this issue when we started getting low readings when the system started up. Upon resetting, the system resets the angle on all three servo arms, this requires a lot of power and would diminish the intensity of the laser emitters. The reduced intensity caused a false-positive situation, in which the system would consider the laser beam interrupted even though it was active. Here is a description of a scenario in which four mice would be trapped: 1.

A mouse would enter through the main door (M on Figure 2) into cell number four. Instantly upon detection by the infra-red sensor, the main door closes, trapping the mouse inside.

The mouse would, after being trapped, begin to wander inside the trap. Eventually, the mouse makes it to cell one (Figure 2), and upon being detected by the sensor inside cell number one, trapdoor (A) is activated.

Once the first mouse is securely trapped in cell one, the main door (M) to cell four opens again.

Another mouse enters through the main door (M) into cell number four. Sensor in cell four detects it and closes the main door (M).

Eventually, the mouse walks to cell number two, and upon entering the sensor detects it and traps it inside by closing trapdoor (B).

Figure 3: Smart Mousetrap in action (3 of the 4 cells implemented at the time of the picture)

5. Testing The trap was tested under many conditions. We ran into problems with the sensibility of the laser sensors to light. Using a 3D Printer, we built small casings to cover the sensors so that they wouldnâ&#x20AC;&#x2122;t receive as much light from the environment. On some situations when the system was receiving too much environmental light (either from the sun or from artificial sources), the sensors would continue to send high readings even with the laser beams were deactivated. This was corrected when we built the casings (however, the trap was meant to be in complete darkness anyways). One important issue we had during the testing of phase two was power. The pairs of laser sender/receiver modules work by shining a laser right into the receiver, which continues to send a highpower signal to the Arduino through the pin itâ&#x20AC;&#x2122;s connected to until this laser beam is interrupted. However, in some cases the laser beam was shining right into the receiver and it would behave as if it had been interrupted. It took a lot of testing to figure out that these were cases in which not enough power was

being supplied to keep the lasers at the right intensity. This meant that the intensity of the laser beam was too weak for the receiver to detect it. Ultimately, we solved this issue by separating the power supply of the laser beams from the power supply of the receivers and motors, using a 12V AC connection.

6. Discussion and Future Work Our prototype in phase two proved to be a very reliable system for trapping multiple mice at a time. But also, we realized that the system could be used for other purposes in addition to trapping mice. Any creature big enough to fit into some trap and capable of interrupting a laser beam could be trapped. Not only for pest control purposes, but also in situations in which someone is performing research in the wilderness and wants to capture a living specimen without harming it in anyway. Servo motors are a rather harmless way of capturing living animals quickly and effectively. We also want to consolidate the trap to be smaller and not need to receive power from an electrical outlet but rather from a battery supply. This will allow the trap to work in outdoor conditions without issue.

7. Conclusion The future of pest control will be more sophisticated and efficient to the degree that we experiment and dare to break barriers with the new technology that is becoming available through the Internet-of-Things revolution. The future is connectivity, and with connectivity we will discover more efficient ways of solving everyday problems by improving current solutions or creating new ones. Practices as ordinary as simply catching a mouse will be changed completely by the arrival of new technologies, showing us a life that we thought only existed in science fiction.

Acknowledgments The authors are thankful to the editorial committee and reviewers for their comments and suggestions, which improved the quality of the paper.

8. References [1] Specialty Consultants Research: U.S. Structural Pest Control Market to Reach $10B in 2020. (2016, April 10). Retrieved May 03, 2019, from https://www.pctonline.com/article/ sc-research-pest-control-marketreport/

[2] How Fast Do Mice Multiply in Your Home? (2018, May 21). Retrieved March 3, 2019, from https://preventivepesthouston.com/ 2018/05/21/mouse-math-how-fastcan-mice-multiply-in-your-home/ [3] Bodkin, Henry. House Mice Carry Deadly Antibiotic-Resistant Diseases, Scientists Warn. 17 Apr. 2018, www.telegraph.co.uk/science/2018/ 04/17/house-mice-carry-deadlydrug-resistant-diseases-scientistswarn/. Accessed 3 Feb. 2019.

Phylogenetic study of the spider community of Simpson Park in Miami, Florida through DNA barcoding Anette Sanz Department of Liberal Arts and Sciences Miami Dade College, Hialeah, Florida Abstract The purpose of this study is to identify the population of the different spider species in Simpson Park of Miami, Florida, as well as to standardize the DNA barcoding of the collected spiders. DNA barcoding was used to supplement the identification of specimens based on morphological similarities. This method, based on DNA sequencing and bioinformatics, demonstrated that of all the universal primers used in our research, the most effective for DNA barcoding of the observed spider community are the forward primer LCO1490 and the reverse primers HCO2198 and HCO-700ME. Fifty spiders from Simpson Park were collected. Only 15 specimens yielded a sufficient DNA concentration for Polymerase Chain Reaction (PCR) after the DNA was extracted and all of them could be effectively barcoded after being selected based on the agarose gel electrophoresis banding results. The barcoded specimens were then introduced into NCBI through the Basic Local Alignment Search Tool (BLAST), which uses various forms of statistical analysis to compare DNA input sequences with a database. The sequences were then used to build the phylogenetic tree using Mega X software, which allowed us to arrive to the conclusion that Argyrodes sp. is the common ancestor of the rest of specimens that were barcoded. Crossopriza, Leiobunum, Verrucosa, Mangora, Gasteracantha, and Nephila are more closely related themselves, and Leucauge and Mecynogea are more closely related themselves than to the first mentioned six genera. Keywords: DNA barcoding, Phylogenetic study, Spider community, Morphological classification 1 Introduction The order Araneae is one of the most diverse arthropod orders. There are thousands of different species of spiders known, and this is considered only a percentage of the real number. Through decades, taxonomists have been able to identify around 50,000 species and more than 100 families of spiders (Crawford, 2015). However, the morphological identification of spider specimens remains complicated. The main limitations presented by traditional methods of specimen classification based on morphological characteristics fall in aspects such as the lack of accurate ways to study non-adult and adult specimen differences, for instance. Additionally, the clear sexual dimorphism - differentiation of male and female structures - makes it extremely difficult and sometimes impossible to provide adequate characterization, and subsequent taxonomic classification of spiders (Barrett & Hebert, 2005). All of this without taking into consideration other anatomical characteristics that the spiders share with other arachnids, and causes the scientific community to be in disagreement when trying to assign them a genus and a specific epithet. While a certain degree of classification can be achieved based on the specimensâ&#x20AC;&#x2122; morphological characteristics, the most accurate species classification can be achieved through the comparison of nucleotide sequences to a database using NCBI-BLAST. The identification of species using their DNA sequence is an actual success in phylogenetics. For this purpose, mitochondrial DNA (mtDNA), specifically that from the Cytochrome Oxidase subunit 1 (COI) has been found is appropriate for barcoding, and it is the most universally accepted and employed in this field. COI is a gene that is present in all spider species with sufficient variation for species differentiation (Barrett & Hebert, 2005).

28 For this purpose, DNA barcoding and taxonomy combined have the potential of contributing to the conservation of the biodiversity in a realistic timeframe, while overcoming the disadvantages of both practices. A DNA barcode is obtained by extracting DNA from the cells of the specimen; amplifying the genetic material through a polymerase chain reaction (PCR) to further sequence the samples that exhibit a band in the region between 500 and 750 base pairs on the agarose gel that corresponds to the mitochondrial gene Cytochrome C Oxidase subunit I (COI). The sequence of the COI gene contains approximately 650 base pairs collected from amplification and is the region of the genome that allows to better classify species based on their sequence arrangement (Bork, 2015). This mitochondrial region is valuable in barcoding since animal cells carry this gene. COI barcodes split enough to permit identification between highly related species. The sequences are aligned using the bioinformatics tools to not only compare across specimens but also to establish their phylogenetic variation. This paper has the aim of exposing methods utilized by our laboratory group at Miami Dade College, to taxonomically classify and study the phylogenetic relationships and genetic variation among the spiders collected at Simpson Park Hammock, Miami. 2 Methods and materials 2.1 Specimen Acquisition and Storage A field exploration was made of the Simpson Park in Miami, Florida to collect the specimens. Fifty samples were collected in 50-mL conical tubes and preserved on ice for transport. Once the specimens were transported to the lab, the tubes were labeled with numbers that ranged from 1-50 and names were given according to the Spiders of Florida guide (Bugh, 2018). A solution of 70 % ethanol was added to the tubes, and the specimens were refrigerated. Afterward, the specimens were removed from the solution, washed twice with deionized water and allowed to dry for at least 2 hours at 30Ë&#x161;C. The specimens were then placed under dissection dishes and viewed under a stereoscope in order to identify various morphological characteristics and reaffirm the name given when compared to the images found in the guide and Florida dichotomous keys. The presumed species names were grouped in a manner to be used as a database to track the different samples through the different protocols during the barcoding process.

Figure 1. Simpson Park of Miami, Florida

Figure 2. Students collecting specimens

2.2 DNA Extraction and Concentration Measurement The dried specimens were first placed into ZR Bashing Bead Lysis Tubes (Zymo ResearchTM) labeled with the assigned specimen number. Smaller specimens were placed directly into the tubes while the larger specimens had their legs removed before placing them into the tubes. To extract the mitochondrial DNA, the Quick-DNATM Tissue/Insect Miniprep Kit from Zymo Research was used following the manufacturerâ&#x20AC;&#x2122;s instructions.

29 Once the DNA was obtained, 2 µL of the DNA solution of every DNA extract were used to measure DNA concentration with a Nanodrop Spectrophotometer ND-1000, and the absorbance at 230 nm and the 260/280 ratio were recorded for every sample. 2.3 DNA Cleaning and Concentration Samples with insufficient DNA concentration (less than 50 ng/µL) or showing contamination by the reagents used in the extraction process (260/280 ratio < 1.8) were concentrated and purified using a DNA Clean & ConcentratorTM kit from Zymo Research. The DNA concentration was measured again to confirm that the final product had a higher concentration and was freed from the residual reagents. 2.4 PCR To design the primer combinations and order the primers that were further used for PCR and sequencing, the National Center for Biotechnology Information (NCBI) Primer Designing tool was used. The forward and reverse primer sequences were taken from the literature (Bork, 2015) and compared to the DNA sequences of the specimens that were morphologically classified. The purpose of using a combination of primers was to select the ones that could provide better results in the PCR and sequencing process. The primers were rehydrated and a 1/10 dilution was added to the DNA in a mixture consisting of 1 µL sample DNA, 13 µL of PCR master mix (GoTaq®Green Master Mix, Promega), 0.5 µL of each primer and 10 µL of nuclease-free water for a total of 25 µL of PCR mixture per sample. The tubes were placed in a thermal cycler SimpliAmpTM, applied biosystems for approximately 3 hours. The primers added in the aforementioned manner were combined in the arrangement presented in Table I. Combinations 1 and 2 have been previously demonstrated to reliably amplify DNA. Primer combination Combination# 1 Combination# 2 Combination# 3 Combination# 4 Combination# 5 Combination# 6

Forward LCO 1409 LCO 1409 Chelicerate forward LCO 1409 LCO 1409 Lepidoptera forward

Reverse HCO 2198 HCO-700 ME Chelicerate reverse 1 Chelicerate reverse 1 Chelicerate reverse 2 Lepidoptera reverse

Table I. Primer combinations for PCR and DNA sequencing

The PCR thermal cycler program consisted of 1 cycle of 1 min at 94˚C (initial denaturation), followed by 35 cycles of 45 sec at 94˚C, 45 sec at 48˚C, 30 sec at 72˚C, and final extension of 1 cycle of 5 min at 72˚C. The samples were then kept at 4˚C until taken out of the thermal cycler to perform the DNA gel electrophoresis. Once the DNA was amplified, 12 µL of the amplified DNA were added to a DNA electrophoresis agarose gel. The gels were visualized and photographed under ultraviolet light in a UVP transilluminator chamber. After the bands were visualized, the samples that exhibit band patterns in the region of interest were selected and sent to Eurofins Genomics for sequencing. 2.5 Phylogenetic Tree Construction

30 The resulting sequences were introduced into the NCBI database and compared using the BLAST analysis tool. The software Mega X was used to align the sequences and construct the phylogenetic tree.

3 Results and Analysis 3.1 DNA Extraction and Concentration After the DNA was extracted as per the instructions, only 10 out of the 50 samples were able to yield concentrations of 50 ng/ÂľL and above. The DNA was then cleaned and concentrated, and 12 more samples were used for PCR. Such success indicates that the methods for increasing DNA concentration, mainly repeated DNA elution and dry centrifugation, were able to reliably increase DNA concentrations to levels suitable for PCR and sequencing. As represented on Table I, primer combination 2 was the most effective primer combination for the barcoding of spiders based on the COI gene because all of the samples displayed banding patterns in the region between 500 and 750 bp of the 1 Kb DNA ladder. Primer combination 1 was still able to demonstrate banding patterns for almost all samples. The reverse primer HCO-700ME was as equally effective at amplifying DNA as the reverse primer HCO2198 as indicated by Figures 5 and 6.

Figure 3. Samples 1 through 11. Primer combination 1

Figure 4. Samples 12 through 22. Primer combination 1

Figure 5. Samples 23 through 33. Primer combination 2

Figure 6. Samples 34 through 44. Primer combination 2

The figures above demonstrate the resulting banding patterns of DNA amplified with the PCR primer combinations 1 and 2. Samples 37 through 44 were not successfully amplified and thus did not display bands (figure 6). The primer combination 3 was used in this case (Table I). The primer combinations 4, 5, and 6 showed banding patterns for just a few samples, which made us select the combinations 1 and 2 as the most appropriate to perform DNA sequencing. It is important to note that no signs of DNA

31 degradation were observed which is an indication of proper specimen storage conditions and manipulation.

3.3 Phylogenetic Tree Analysis A phylogenetic or evolutionary tree is a graphic way to represent the evolutionary relationships among groups of organisms or taxa. The nodes of the tree represent the common ancestors, and the branches represent the groups of descendent taxa, commonly species. After the DNA sequences for each specimen were obtained, the aligned sequences were analyzed using the NCBI-BLAST database to find the specimen identity. Most of the specimens were found to belong to the species Gasteracantha cancriformis and Leucage argyra, which were found to be the most prevalent species inhabiting the park, with a relative abundance of 48% and 24% respectively. Figure 7 represents the phylogenetic tree built using the Mega X software, which helped us to establish the phylogenetic relationships among the studied specimens. The results demonstrated that the members of the genus Argyrodes are the common ancestors of all the spider species found in the park. The phylogenetic tree also shows that Crossopriza, Leiobunum, Verrucosa, Mangora, Gasteracantha, and Nephila are more closely related themselves, and Leucauge and Mecynogea are more closely related than the first mentioned six genera.

Figure 7. The evolutionary history was inferred using the Maximum Parsimony method

4 Conclusions In this project, 50 spider specimens from Simpson Park, Miami were collected and morphologically classified. Using DNA barcoding, it was possible to identify the species to which the samples belonged to as well as select the best of various primers used to amplify and sequence the DNA. Many samples were unable to be barcoded, and so the relative abundance of spiders was calculated based on morphological characteristics. It was also demonstrated that storing the specimens in a solution of 70% ethanol allowed for later extraction of genetic material that prevented DNA degradation and did not interfere with the PCR and sequencing reactions. The most effective primer combinations for arachnid DNA barcoding are the

32 forward primer LCO1490 and the reverse primers HCO2198 and HCO -700ME, which allowed to detect the target gene for the Cytochrome Oxidase subunit 1 of the mitochondrial DNA. The results of the phylogenetic tree analysis demonstrated that the members of the genus Argyrodes are the common ancestors to all the spider species collected in the park. The phylogenetic tree also shows that Crossopriza, Leiobunum, Verrucosa, Mangora, Gasteracantha, and Nephila are more closely related themselves, and Leucauge and Mecynogea are more closely related themselves than to the first mentioned six genera. The study of the spider population of the Simpson Park as a representation of the spider biodiversity of Miami, Florida has been done for the first time since there are no records found about phylogenetic research in this area. This study adds to our knowledge experience in the field of taxonomy and systematics, and in the use of the most recent protocols applied from the specimen collection and preservation to the DNA sequencing and bioinformatics. The students exposed to this type of research not only develop skills in the DNA manipulation techniques but also increase their sensitivity to the spider’s habitat destruction when the preservation of the biodiversity is a challenge.

Acknowledgments The author is thankful to the editorial committee and reviewers for their comments and suggestions, which improved the quality of the paper. Also, the author is thankful to the Chair of Liberal Arts and Sciences, Dr. Caridad Castro, the SSAS Grant Director, Ms. Laura Iglesias, and to Miami Dade College for the opportunity to serve the LAS Department and Hialeah Campus.

References Barrett, R. D., Hebert, P. D. (2005). Identifying spiders through DNA barcodes. Retrieved April 19, 2019, from https://www.nrcresearchpress.com/action/showCitFormats?doi=10.1139/z05- 024 Bork, R. J. (2015). Primer efficacy in the DNA barcoding of spiders (Doctoral dissertation, University of Northern Iowa) Breitling, R. (2017). Public DNA barcoding data resolve the status of the genus Arboricaria (Araneae: Gnaphosidae). Arachnologische Mitteilungen. 54 (54): 24-27. Retrieved from: https://www.researchgate.net/publication/318792137Public_DNA_barcoding_data_resolve_the_s tatus_of_the_genus_Arboricaria_Araneae_Gnaphosidae Bugh, V. G. (2018). Spiders of Florida. A Guide to Common and Notable Species. Pamphlet. Coddington, J. (2005). “Phylogeny and Classification of Spiders”. American Arachnological Society Crawford, R. (2015). “Myth: spiders are easy to identify”. Retrieved March 26, 2019, from https://www.burkemuseum.org/blog/myth-spiders-are-easy-identify Dhaliwal, A. (2019, February 18). DNA Extraction and Purification. Retrieved March 26, 2019, from https://www.labome.com/method/DNAExtraction-and-Purification.html DNA Clean & ConcentratorTM Kit Protocol. (n.d). Zymo Research Corporation. Retrieved on March, 2019 from: https://files.zymoresearch.com/protocols/_D4003T_D4003_D4004_D4013_D4014_DNA_Clean_ Concentrator_-5_ver_1_2_1_LKN-SW__1.pdf

33 Doorenweerd, C., Beentjes, K. (2012, January). Extensive guidelines for preserving specimen or tissue for later DNA work[PDF] GoTaq® Green Master Mix (M712) Product Information. (2018, April). Retrieved April 22, 2019, from https://www.promega.com/resources/protocols/product-informationsheets/g/gotaq-green-mastermix-m712-protocol Insects of Florida. (n.d.). Retrieved March 26, 2019, from https://www.insectidentification.org/insectsbystate.asp?thisState=Florida Kumar S., Stecher G., Li M., Knyaz C., and Tamura K. (2018). MEGA X: Molecular Evolutionary Genetics Analysis across computing platforms. Molecular Biology and Evolution 35:1547-1549 Mayfield, T. (n.d.). Distinguishing between ethanol and isopropanol in natural history collection fluid storage [PDF]. Chicago, IL: Field Museum, Division of Amphibians and Reptiles Nei M. and Kumar S. (2000). Molecular Evolution and Phylogenetics. Oxford University Press, New York Quick-DNATM Tissue/Insect Miniprep Kit Protocol. (n.d.). Zymo Research Corporation. Retrieved on March, 2019 from www.zymoresearch.com/products/quick-dna-tissue-insect-miniprep-kit Ramakers C, Ruijter JM, Lekanne Deprez RH, Moorman AFM. (2002). “Assumption-free analysis of quantitative real-time polymerase chain reaction (PCR) data”. Neuroscience Letters. 339: 62-66 Wegner, G. S., Ph.D., BCE. (2011, February). Spider Identification Guide. Retrieved March 26, 2019, from https://ipminstitute.org/wpcontent/uploads/2016/06/Spider-Guide-Wegner-BASF-Revised12-2-14.pdf Wilson, John. (2012). DNA barcodes for insects. Methods in molecular biology (Clifton, N.J.). 858. 1746. 10.1007/978-1-61779-591-6 3 Zietkiewicz E, Rafalski A, Labuda D. (1993). “Genome Fingerprinting by Simple Sequence Repeat (SSR)-Anchored Polymerase Chain Reaction Amplification”. Genomics, 20. 176-183

A Statistical-based Analysis on the Effect of the Factors Underlining the Structure of the TIMSS Instrument for Measuring Perception of Students toward Mathematics and their Teachers and Gender Dr. Nelson De La Rosa, PhD Professor of Mathematics and Statistics Department of Mathematics, Kendall Campus E-mail: ndelaro1@mdc.edu Miami Dade College, Kendall Campus Abstract The Trends in International Mathematics and Science Study (TIMSS) is an international assessment of mathematics and science that stores information of student achievement as well as data pertaining to the context in which students learn mathematics and science. Research studies have indicated that a comprehensive assessment of student success in mathematics needs to expand the scope of analysis beyond examining performance itself. Attitudes and beliefs have been identified as key factors in student success. Furthermore, research studies have claimed that knowing the extent of the effect of student attitudes on their performance allows educators to rethink teaching and learning. This study examined the structure of the instrument used in the TIMSS to assess the perception of students toward mathematics and their teachers. It also analyzed the relationship between the factors that emerged from the structure of the instrument and gender. Exploratory factor analysis (EFA) was used to extract 4 factors from the TIMSS measure. A one-way MANOVA assessed the relationship between the factors that emerged from the structure of the instrument and gender. This analysis found that that female students achieved higher than males in two of the factors (apprehension and engagement). A reflection on the findings of the study is provided. A case is made for further studies.

Keywords: TIMSS, Attitudes toward mathematics, Exploratory Factor Analysis, MANOVA. Introduction The Trends in International Mathematics and Science Study (TIMSS) is an international assessment of mathematics and science that has served as a reference to study cross-national achievement data (Wang, 2011). The TIMSS is a sophisticated repository of data that stores information of student achievement. The TIMSS also includes data pertaining to the context in which students learn mathematics and science (Wang, 2011). Research studies have indicated that a comprehensive assessment of student success in mathematics needs to expand the scope of analysis beyond examining performance itself. Attitudes and beliefs have been identified as key factors in student success (Ma & Kishor, 1997). Furthermore, research studies have claimed that knowing the extent of the effect of student

attitudes on their performance allows educators to rethink teaching and learning. Seider, Gilbert, Novick, and Gomez (2013) and Wang (2011) found that student attitudes are influenced by their experiences in the classroom. Guven and Cabakcor (2013) and Lusby (2012) concluded that both student beliefs regarding mathematics and classroom environment are key factors in shaping student attitudes toward school and success in the subject. Along with these findings, gender has been identified as a potential predictor of attitudes and beliefs toward mathematics (Gunderson et al., 2012; Hyde & Mertz, 2009). Gunderson et al. (2012) claimed that gender stereotypes may exacerbate learner anxiety and frustration and reaffirm gender differences in mathematics. Gunderson and colleagues suggested that the idea persists that anxiety and frustration affect female students at a higher rate than their male counterparts. According to Gunderson et al. (2012), it is not surprising that female students are less driven to pursue mathematics and science careers than male students since female students have been frequently stigmatized as poor mathematics performers. Hyde and Mertz (2009) denied that girls are less capable to deal with mathematics than boys and attributed any difference in attitude or performance to the influence of external factors rather than to the effect of gender. In an analysis on attitude toward mathematics, Seider and colleagues found that high-level achievement requires long term commitment and responsibility independent of gender. Based on the framework established by Seider et al. (2013) and Wang (2011), this study examined the structure of the instrument used in the TIMSS to assess the perception of students toward mathematics and their teachers. It also analyzed the relationship between the factors that emerged from the structure of the instrument and gender. To this end, the following overarching research questions were used: Q1. Is there more than one reliable and interpretable factor underlying the structure of the TIMSS instrument used to assess perceptions toward mathematics and their teachers? Q2. Is there a relationship between gender and the factors underlying the structure of the TIMSS instrument used to assess perceptions toward mathematics and their teachers? Methodology Data Source and Participants This study used 9798 of the cases of eighth graders who participated in the TIMSS in mathematics from nine states of the United States. Four thousand nine hundred and ninety-four (about 51%) of the cases were female and 4803 (about 49%) were male. Most of the students were around 14 years of age (M = 14.23, SD = 0.48, range = 12 to 18). Data collection was conducted by the National Assessment of Educational Progress (NAEP) in conjunction with the collaboration of personnel from the schools, school districts, and department of education of the participant states. The data collection was conducted while protecting the identity of the participants and confidentiality of the responses. Data used in the analysis included the responses of students who completed the United States student questionnaire, an adaptation of the general/integrated science version of the TIMSS student questionnaire. The questionnaire consisted of non-academic items related to mathematics that aimed to collect both factual and

opinionated responses of students. The analysis of this study focused on items that measured mathematics perceptions of students and their teachers. Procedure The study used exploratory factor analysis (EFA) with oblique and orthogonal rotations to assess the dimensionality of a scale containing 25 items in the questionnaire. The initial stage of this analysis sought to extract factors to make decisions on the number of factors underlying the measures. Further analyses based on item contribution to the factors were conducted to arrive at a plausible solution that would appropriately describe the relationship data-instrument. A oneway MANOVA was conducted to assess the effect of gender on the factors obtained from the factor analyses. All analyses were conducted in SPSS. An alpha level of .05 was used for all statistical tests except in analysis that required stringent values of significance.

Results

Research Question 1 Exploratory Factor Analysis Exploratory factor analysis (EFA) using principal axis factoring (PAF) extraction and the orthogonal rotation of Promax of psychometric items administered in the 2011 TIMSS U.S. questionnaire was performed on the data of 9798 cases of 8th grade students. The analysis focused on grouping the items in dimensions that would reflect the themes addressed in the instrument.

Extraction of the Factors Twenty-five items addressing perception of students toward mathematics and their teachers were analyzed. Prior to running the exploratory factor analysis (EFA), data were screened by examining descriptive statistics on each item, interim correlations, and possible univariate and multivariate assumption violations. From this initial assessment, all variables were found to be interval-like, the distribution of the variable pairs appeared to be bivariate normally distributed, and case independence was warranted. The large sample size warranted the ratio variables-tocases as adequate. Additionally, analysis of the correlations between items of the scale was assessed. Item inter-correlations were positive and significant and ranged from weak to strong without defining a clear pattern of items clustering. A preliminary analysis on the reliability of the measure via Cronbach Alpha index achieved a value of 0.912. A further attempt to reach a higher value of the reliability index was conducted. This effort was accomplished by eliminating item 21, which in turns produced a Cronbach Alpha value of 0.924. An inspection of the wording of the items in the scales hypothesized that a solution with at least four dimensions would be likely to emerge as the description of the items involved different themes. Prior to conducting the EFA, the suitability of the data for this analysis was

examined (Field, 2009; Hair, Black, Babin, & Anderson, 2010; Meyers, Gamst, & Guarino, 2015). The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.94 for the mathematics perceptions scale and 0.781 for the teacher’s scale, indicating that the data was suitable for factor analysis. Similarly, Barlett’s test was significant for both scales (  2 Mathematics = 119895.93,

 2Teacher = 15191.71, p < .001), granting sufficient correlation between the variables to proceed with the analysis. An initial EFA was carried on the mathematics perception and the teacher’s effect scale without imposing a restriction on the number of factors. Results of this assessment are presented in Table 1. In this analysis, 56% of the total variance of the mathematics perception scale was explained and it produced a poorly defined 4-factor solution. A one-factor solution from the teacher’s scale just accounted for 44% of the variance of perceptions of students. The KaiserGuttman retention criterion of eigenvalues greater than 1 was used as a reference for factor extraction and the scree plot shown in Figure 1 confirmed the structure of both solutions. As noticed from Table 1, the initial solution posed several challenges. There were items with factor loading and communalities lower than 0.7 and 0.5 respectively. Additionally, there was cross loading of items. Field (2009), Hair et al. (2010), and Meyers et al. (2015) recommend retaining items with values of communalities and loading closer to 0.5 and 0.7 respectively, as well as avoid cross-loading of less than 0.2.

Table 1 Summary of the initial solution from an exploratory factor analysis on the items of the mathematics perception and teacherâ&#x20AC;&#x2122;s scales (N =9798). Mathematics Teacherâ&#x20AC;&#x2122;s scale scale Item F-1 F-2 F-3 F-4 F-5 Communality Item 20 .859 .700 Item 13 .858 .604 Item 14 .781 .692 Item 16 .734 .515 Item 15 .432 .423 .632 Item 12 410 .397 .566 Item 5 .848 .803 Item 1 .831 .782 Item 3 .820 .590 Item 2 .703 .523 Item 4 .678 .538 Item 24 .793 .523 Item 23 .753 .462 Item 22 .591 .403 Item 6 .540 .348 Item 25 .316 .322 .396 Item 18 .852 .568 Item 19 .759 .512 Item 17 .304 .495 .564 Item 10 .839 .704 Item 11 .767 .589 Item 9 .734 .538 Item 7 .530 .281 Item 8 0.87 Eigen-value 7.983 2.498 1.320 1.144 2.658 % of 37.914 10.064 4.582 3.532 43.988 variance

Figure 1. Scree plot of the initial solution from an exploratory factor analysis on the items of the mathematics perception and teacherâ&#x20AC;&#x2122;s scales.

Based on Field (2009), Hair et al. (2010), and Meyers et al. (2015) recommendations, a comprehensive approach of analysis sought to achieve a factor solution that would retain as many items as possible and that would also consider both item wording and themes intended to measure. Subsequent EFAs were conducted by fixing the number of factors, eliminating items one at a time, and implementing several rotation strategies. An intermediate 3-factor solution still produced not clear pattern in the mathematics perceptions scale including problems such as low values of factor loading and communalities as well as inter-factor cross loading (Table 2, Figure 2). As a result, items with poor contribution in terms of loading and communalities (or which statement was too general) were suppressed.

Table 2 Summary of the intermediate solution from an exploratory factor analysis on the items of the mathematics perception and teacherâ&#x20AC;&#x2122;s scales (N = 9798). Mathematics Perception Teacherâ&#x20AC;&#x2122;s effect scale scale Item F-1 F-2 F-3 F-4 Communality Item 13 .872 .619 Item 14 .785 .688 Item 20 .783 .664 Item 15 .767 .665 Item 16 .749 .545 Item 12 .746 .621 Item 17 .705 .587 Item 18 .574 .429 Item 19 .532 .400 Item 3 .894 .697 Item 2 .814 .628 Item 5 .794 .800 Item 1 .787 .783 Item 4 .763 .614 Item 25 .411 .343 .520 Item 23 .832 .613 Item 24 .781 .585 Item 22 642 .471 Item 6 .628 .445 Item 10 .839 .704 Item 11 .767 .589 Item 9 .734 .538 Item 7 .530 .281 Item 8 0.87 Eigen-value 7.763 2.234 1.318 3.532 % of 40.860 11.756 6.939 43.988 variance

Figure 2. Scree plot of the intermediate solution from an exploratory factor analysis on the items of the mathematics perception and teacher’s scales A final 3-factor solution meeting the criteria above established was achieved for the mathematics perception scale and a one-factor solution for the teacher’s scale (Table 3). As for previous analysis, the factor extraction followed the Kaiser-Guttman retention criterion of eigenvalues greater than 1, as shown in the scree plot (Figure 3). The final solution cumulatively accounted for 63.156% and 60.962% of the total variance of the mathematics perception and teacher’s scales respectively. Three factors were extracted from the mathematics perceptions extraction. This solution included items 13, 20, 14, and 16 in the first factor. Items 5, 1, and 4 composed the second factor. The last factor included items 23 and 24. Items 10, 11, and 9 compose the factor from the teacher’s scale. Such factor composition led to naming factor 1 as Apprehension, factor 2 as Engagement, factor 3 as Career Goal, and the factor from the teacher’s scale as Classroom Climate. A primary goal adopted in this study was to achieve the highest explained variance while retaining items whose contributions were significantly captured by the factors to come up with a plausible structure. Therefore, the results aforementioned strongly suggested the appropriateness of a 4-factor solution. The internal consistency of each factor was also assessed via Cronbach’s alpha. Scores of this index showed that factors 1, 2, 3 reached good values and acceptable for factor 4 (Table 3).

Table 3 Summary of the final solution from an exploratory factor analysis

results on the items of the mathematics perception and teacher’s effect measure (N = 9798). Mathematics Perception Teacher’s scale scale Item F-1 F-2 F-3 F-4 Communality Item 13 .846 Item 20 .815 Item 14 .765 Item 16 .712 Item 5 .901 Item 1 .887 Item 4 .767 Item 23 .739 Item 24 .731 Item 10 .854 .730 Item 11 .793 .629 Item 9 ..685 .469 Eigen-value 4.597 1.697 1.148 2.206 % of 42.692 12.944 7.520 60.962 variance Cronbach’s 0.867 0.877 .819 0.693 Alpha

Figure 1. Scree plot of the final solution from an exploratory factor analysis on the items of the mathematics perception and teacher’s scales.

Research Question 2

A one-way MANOVA assessed the relationship between the four dimensions obtained from the factor analyses and gender. Four composite variables representing the factors extracted were computed following the structure shown in Table 3. These variables were created as summated scales by adding the score of each of the items loading in the corresponding factor and they served as the dependent variables. The technique of using summated variables as the form of creating composite variables is recommended when multivariable analysis is further required (Hair et al., 2010). The independent variable in this analysis was the gender of students. An exploration of the correlation coefficients was conducted among the four mathematics perception dimensions, prior to run the one-way MANOVA. All correlations were statistically significant, despite some of these relationships were weak (Table 4). Table 4 Correlation coefficients of Apprehension, Engagement, Career Goal, and Classroom Climate 1 2 3 4 1. Apprehension 1 .517** 1 2. Engagement ** .135 .292** 1 3. Career Goal ** ** .288 .594 .224** 1 4. Climate **. Correlation is significant at the 0.01 level (2-tailed). Results from the one-way MANOVA are shown in Table 5. Bartlett’s test of sphericity was statistically significant (  2 = 18577.24, p<.001), suggesting that there was sufficient correlation between dependent variables. Due to a statistically significant (p < .01) Box’s test of equality of covariance-covariance matrices, indicating that the dependent variable covariance matrices were not equal across the levels of the independent variable, a Pillai’s trace was used to evaluate the multivariate effect. Using Pillai’s trace as the criterion, the composite dependent variate was significantly influenced by gender, Pillai’s trace = .006, F (4, 9793) = 15.63, p < .01.

Table 5 One- way MANOVAs on Apprehension, Engagement, Career Goal, and Classroom Climate across Gender. Gender Box's Test Bartlett's Test of Multivariate Tests Sphericity Female Male F (10,457224539) p χ²(9) P Pillais’s p Partial 4994

4804

42.30

<001

18577

<.001

.006

<.001

2 .006

Univariate ANOVAs were conducted on each dependent measure separately to determine the locus of the statistically significant multivariate effect (Table 6). Two statistically significant univariate effect were found, one associated with apprehension F (1, 9796) = 43.62, p < .001, 2 = .004 and the other one with engagement F (1, 9796) = 32.94, p < .001, 2 = .003. This finding indicated that female students achieved higher scores in apprehension [M = 9.11, SD = 3.73] and engagement [M = 6.79, SD = 2.72] than male students [M = 8.26, SD = 3.56; M = 6.47, SD = 2.70]. Table 6 One-way ANOVAs on Apprehension, Engagement, Career Goal, and Classroom Climate across Gender. Levene’s ANOVAs Female Male 2 F(1,9796) p F(1,9796) p M SD M SD  Apprehension 20.01 <.001 43.62 <.001 .004 9.11 3.73 8.62 3.55 Engagement 2.39 .122 32.94 <.001 .003 6.78 2.72 6.47 2.70 Career Goal 5.34 <.021 2.01 <.151 .000 2.84 1.22 2.82 1.17 Classroom .422 <516 .359 <.549 .000 6.30 1.22 6.28 2.45 Climate

Discussion and Conclusions The present multipurpose study aimed to analyze the structure of the mathematics-related psychometric items of the TIMSS. Further, it assessed the relationship between gender and those factors that emerged from the psychometric measure of the TIMSS. Rather than analyzing the structure of the measure of the TIMSS, previous studies have focused on addressing achievement. Consequently, this study departed from the common trend of research on cross sectional data and represents a singular attempt to explore the influence of non-cognitive measures. The results of this study corroborated the assumptions that the mathematics perceptions would produce four factors, three from the perceptions toward mathematics and one from the teacher’s measure. The factors represented the themes of apprehension and engagement to perform mathematics tasks, career goal, and classroom climate. These findings are consistent with Ma and Kishor’s (1997) thesis regarding the different representations that can emerge from mathematics attitudes. Apprehension and engagement, two opposed yet related themes, came up as relevant expressions of attitude from the exploratory factor analyses. Such an outcome stresses that apprehension reduces the capacity for students to properly function and achieve (Lusby, 2012), as well as that engagement is vital to foster efficacy (Seider et al., 2013). Additionally, career goal orientation proved to be of concern for students in the study sample. Previous studies have linked high achievement in mathematics to positive attitude toward mathematics (Lusby, 2012; Seider et al., 2013; Zuffianὀ et al., 2013). Therefore, it is expected that positive attitudes are somewhat related to high expectations in a student’s future.

The finding with respect to the teachers’ measure was also consistent with the initial assumption in this respect. Research reports that teachers are moderators of students’ attitudes. According to Özdemir and Pape (2013) teachers’ actions play a crucial role in developing internal encouragement (student’s motivation from teacher’s action) and external encouragement (feedback students receive from the teacher). Özdemir & Pape further claimed that teachers may strength or reduce students’ beliefs in their skills to succeed in mathematics. Teacher role is crucial in identifying when a student is operating in isolation as well as in guiding them to overcome setbacks (Singh et al., 2002). Bandura (1994) established that external factors such as schools are operational agents that influence the attitude of students to mathematics. Mixed findings arose from the analysis of the effect of gender on the dimensions extracted from the TIMSS measures. On one hand, the belief that female students score differently than male students on mathematics perceptions and teachers’ scales was confirmed, debunking the claim that persists in schools and in society (Gunderson et al., 2012; Hyde & Mertz, 2009). It appears that female students in this sample experienced more apprehension than their male counterparts. However, female students were slightly more efficient and engaged than male students in the mathematics classroom. Both female and male students showed similar judgment regarding the effect of teachers. Despite that the size of the effect was low in all these categories, the findings in this study indicate that overall girls are not less mathematics-driven than boys. This result may be interpreted as a consequence of the effort schools have been deploying to encourage female students to improve their motivation and attitude to mathematics. Results from this study indicate the need of replicating studies like this at different academic levels. Research in mathematics has striven to find the roots of low achievement in mathematics. Despite efforts deployed towards this endeavor, there is still much left to understand about how students build their beliefs toward mathematics. Low student achievement in mathematics is a reason for concern at all academic levels (Peterson et al., 2011). Findings in this study were limited to the middle school context. Most studies that have addressed the role of internal factors and external factors on mathematics using a big sample have been conducted at the middle school level. Colleges and universities require undergraduate students to complete general education courses in mathematics. Therefore, analyzing data from both secondary and postsecondary institutions will provide a meaningful framework for understanding the development of student attitudes and beliefs toward mathematics across academic levels.

APPENDIX Perceptions toward mathematics and teacher items in the TIMSS instrument Item 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Description I enjoy learning mathematics I wish I did not have to study mathematics Mathematics is boring I learn many interesting things in mathematics I like mathematics It is important to do well in mathematics I know what my teacher expects me to do I think of things not related to the lesson My teacher is easy to understand I am interested in what my teacher says My teacher gives me interesting things to do I usually do well in mathematics Mathematics is more difficult for me than for many of my classmates Mathematics is not one of my strengths I learn things quickly in mathematics Mathematics makes me confused and nervous I am good at working out difficult mathematics problems My teacher thinks I can do well in mathematics classes with difficult materials My teacher tells me I am good at mathematics Mathematics is harder for me than any other subject I think learning mathematics will help me in my daily life I need mathematics to learn other school subjects I need to do well in mathematics to get into the college or university of my choice I need to do well in mathematics to get the job I want I would like a job that involves using mathematics

Acknowledgements The author of this study is thankful to the Trends in International Mathematics and Science Study (TIMSS) and the National Center of Education Statistics (NCES) for providing free access to valuable information and data resources available on their website (https://nces.ed.gov/timss/) which were consulted to write this paper. The author is also thankful to the editor-in-chief and reviewer for providing constructive feedback which improved the quality of the paper.

References

Gunderson, E., Ramirez, G., Levine, S., & Beilock, S. (2012). The Role of Parents and Teachers in the Development of Gender-Related Math Attitudes. Sex Roles, 66(3/4), 153-166. Retrieved from http://ehis.ebscohost.com.ezproxy.fiu.edu/ehost/pdfviewer/pdfviewer?vid=5&sid=28f10 e30-2f53-4f60-b528-29f8556f47b3%40sessionmgr13&hid=8 Guven, B., & Cabakcor, B. (2013). Factors influencing mathematical problem solving achievement of seventh grade Turkish students. Learning & Individual Differences, 23131-137. Retrieved from http://www.sciencedirect.com.ezproxy.fiu.edu/science/article/pii/S1041608012001434 Field. A. (2009). Discovering statistics using SPSS. Thousand Oaks, CA: Sage. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall. Hyde, J. S. & Mertz, J. E. (2009). Gender, culture, and mathematics performance. Proceedings of the National Academy of Sciences of the United States of America. Retrieved from https://doi.org/10.1073/pnas.0901265106 Lusby, B. (2012) Increasing Student's Self-efficacy in Mathematics. Rising tide Volume 5. St. Maryâ&#x20AC;&#x2122;s College of Maryland. Retrieved from http://www.smcm.edu/educationstudies/pdf/rising-tide/volume-5/Lusby.pdf Ma, X., & Kishor, N. (1997). Assessing the relationship between attitude toward mathematics and achievement in mathematics: A meta-analysis. Journal for Research in Mathematics Education, 28(1), 26-47. doi:http://dx.doi.org.ezproxy.fiu.edu/10.2307/749662 Meyers, L. S., Gamst, G., & Guarino, A. J. (2015). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage. Ă&#x2013;zdemir, I., & Pape, S. (2013). The Role of Interactions between Student and Classroom Context in Developing Adaptive Self-Efficacy in One Sixth-Grade Mathematics Classroom. School Science & Mathematics, 113(5), 248-258. Retrieved from http://ehis.ebscohost.com.ezproxy.fiu.edu/ehost/detail?vid=6&sid=28f10e30-2f53-4f60b52829f8556f47b3%40sessionmgr13&hid=8&bdata=JnNpdGU9ZWhvc3QtbGl2ZQ%3d%3d #db=eft&AN=87454017

Seider, S., Gilbert, J. K., Novick, S., & Gomez, J. (2013). The Role of Moral and Performance Character Strengths in Predicting Achievement and Conduct among Urban Middle School Students. Teachers College Record Volume 115 Number 8, 2013, p. - Number: 17075. Retrieved from http://www.tcrecord.org/library/ Wang, J. (2011). Re-examining Test Item Issues in the TIMSS Mathematics and Science Assessments. School Science & Mathematics, 111(7), 334-344. Retrieved from http://ehis.ebscohost.com.ezproxy.fiu.edu/ehost/pdfviewer/pdfviewer?vid=3&sid =9a74c217-86eb-4776-984a-0bf4807b643f%40sessionmgr15&hid=6

Statistical Analysis through Opinion Mining in the Trends of Cruise Industry Dr. Lourdes Gonzalez, PhD Professor of Mathematics Department of Mathematics E-mail: lgonzal2@mdc.edu Miami Dade College, Kendall Campus

Dr. Nelson de la Rosa, PhD Professor of Mathematics and Statistics Department of Mathematics E-mail: ndelaro1@mdc.edu Miami Dade College, Kendall Campus

Abstract Cruise companies have made costumers satisfaction their top priority for achieving success. Aimed to promote a culture of learning communities in our students where mathematics and other disciplines converge to analyze current phenomena, a group of STEM students across campuses of the Miami Dade College were mentored on exploring similarities and differences between consumer needs on commercial cruise ships. Data following random sampling selection of 180 different reviews from individuals who had traveled on a cruise ship in the past year was collected from the TripAdvisor website. Opinion mining, a variant of data mining was used to reveal pattern of association. Findings from this study showed differences of consumerâ&#x20AC;&#x2122;s choice criteria at the time of booking commercial and luxury cruise lines. While consumers traveling in either commercial or luxury cruises found great service, great food, and great excursions as key factors for cruise selection, those traveling in luxury cruises were more vocal at the time of expressing negative opinion when their needs regarding quality of service and room conditions did not meet their needs. Keywords: Cruise companies, Opinion mining, Frequency matrices, Singular value decomposition, Word coefficients. Introduction Analysis of costumerâ&#x20AC;&#x2122;s behaviors is a fundamental principle for business organizations to guarantee their successful operation, expansion, and survival, regardless their size. Cruise companies have experienced significant global growth for more than five decades. Both commercial and luxury liners have made costumers satisfaction their top priority for achieving success. This paper reports the findings of the first stage of an ongoing study aimed to promote a culture of learning communities in our students where mathematics and other disciplines converge to analyze current phenomena. Two mathematics faculty from the Kendall Campus of Miami Dade College mentored a group of students pursuing STEM majors from Miami Dade College campuses on exploring similarities and differences between consumer needs on commercial cruise ships. To this end, students and faculty worked closely together in a project that combined complex mathematics tools as well as concepts from social science disciplines.

The sections that follow provide a thorough description on the methodology used to accomplish successfully the first stage of this study. Methodology Pattern of association was carried out using data mining. Data mining is a complex statistical methodology that segments data structure and uses classification, clustering, and forecasting to discover patterns and relationships (Han, Kamber, Pei, & Kaufmann, 2012).. Opinion mining was the modality of data mining used in this study. Opinion mining is related to the necessity of knowing what other people think. It is often linked to sentiment analysis and deals with the computational treatment of opinion, sentiment and subjectivity in texts (Han, Mankad, Gavirneni, & Verma, 2016). Data following random sampling selection of 180 different reviews from individuals who had traveled on a cruise ship in the past year was collected from the TripAdvisor website. Ninety of these reviews were from individuals who traveled on commercial cruise ships while the rest of the reviews came from individuals who traveled in luxury cruise lines. Combined application of tools of Statistics and Linear Algebra were used to handle the big amount of unstructured text data and discover underlying relationships. Concepts showed relationships between words and how they are related. Most frequently used words were analyzed as well. Statistical analysis included text cleaning and pattern extraction. Text cleaning was conducted in Microsoft Excel. Pattern extraction using the techniques of frequency matrices, singular value decomposition, word coefficients, and scree plots were conducted in Statistica. Frequency matrices and the method of Singular Value Decomposition were utilized to extract meaningful words and concepts that were not clearly observed without using latent variables. Steps in the Analysis Concepts and Techniques utilized in data mining included text mining that allowed examining the collection of written resources to generate new information. By identifying facts and relationships, it transformed unstructured (textual) information into structured data for analysis and visualization. A preliminary text cleanup was performed by removing unnecessary information and, spelling errors such as removing (stop) words like “the”, “a”. Combining abbreviations and different grammatical forms of the same words (stemming) facilitated analyzing and possible condensation of synonyms and phrases. Additional cleaning consisted in tokenization that broke texts into words, terms, and other meaningful elements or tokens, generation of a data matrix of documents and words (matrix of frequencies containing the number of times a word occurs in each document). Furthermore, pattern association used transformations to summarize the information extracted. These transformations included: Binary frequencies: f(wf) = 1, for wf > 0

and

0 đ?&#x2018;&#x2013;đ?&#x2018;&#x201C; đ?&#x2018;¤đ?&#x2018;&#x201C;đ?&#x2018;&#x2013;,đ?&#x2018;&#x2014; = 0 đ?&#x2018;&#x2018;đ?&#x2018;&#x201C;(đ?&#x2018;&#x2013;, đ?&#x2018;&#x2014;) = { đ?&#x2018; (1 + log(đ?&#x2018;¤đ?&#x2018;&#x201C;đ?&#x2018;&#x2013;,đ?&#x2018;&#x2014; ))đ?&#x2018;&#x2122;đ?&#x2018;&#x153;đ?&#x2018;&#x201D; đ?&#x2018;&#x2018;đ?&#x2018;&#x201C; đ?&#x2018;&#x2013;đ?&#x2018;&#x201C; đ?&#x2018;¤đ?&#x2018;&#x201C;đ?&#x2018;&#x2013;,đ?&#x2018;&#x2014; â&#x2030;Ľ 1 đ?&#x2018;&#x2013;

Latent Semantic Indexing Latent semantic indexing identified underlying (latent) dimensions (of â&#x20AC;&#x153;meaningâ&#x20AC;?) was used to create a mapping of the word and documents into a common space, computed from the word frequencies or transformed word frequencies. This technique is a common analytic tool for interpreting the â&#x20AC;&#x153;meaningâ&#x20AC;? or â&#x20AC;&#x153;semantic spaceâ&#x20AC;? described by the words extracted from the documents. Singular value decomposition Singular value decomposition served to reduce the overall dimensionality of the input matrix (word-document) to a lower-dimensional space so that the most important dimensions are identified and noisy correlations found in the original data, as well as dimensions with no meaningful variation, are disregarded. This technique is applied in methods used for document retrieval and analysis of word similarity. Once the dimensions were identified, the underlying latent meaning of what is contained in the documents were retained. Results Travelers on commercial cruise ships indicated that quality of their service on the ship and of the food as a top priority for booking a cruise. Additionally, it appears that travelers on luxury cruise ships also remarked food and service on the ship as an important aspect to consider when choosing a cruise, despite they found that these amenities did not quite met their needs. Results of the statistical analyses for each type of cruise are presented below. Commercial Cruises The scree plot showed an inflection point at Concept 6. Therefore, by the elbow method, 6 concepts (vectors) were selected to continue the analysis. For each concept, the words (components) with more relative weight were selected from the table accompanying the scree plot. Singular values 7.0 6.5 6.0

Singular value % explained

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0

2 1

4 3

6 5

8 7

10 12 14 16 18 20 22 24 26 28 30 32 34 36 11 13 15 17 19 21 23 25 27 29 31 33 35 37 Concept

The concepts described by the words that carried the higher weight (greatest absolute valued coefficients) were as follows: Concepts for Commercial Cruises Concept 1: + great + room + good + ship Concept 2: - great-excursions – great-shows – great-service – great-food Concept 3: + great-service + excellent / - great-ship Concept 4: + room Concept 5: + small + food / - dine – cruise Concept 6: + great-service + great-show/ - children – great-excursion As observed, signs of the word coefficients in Concepts 1 were all positive. This finding revealed a co-occurrence of the words “great”, “room”, “good” and ”ship” across the opinions. The signs of the word coefficients in Concept 2 were all negative, indicating co-occurrence as well. Concept 6 showed two groups where “great service” and “great show” appeared together. Besides, “children” and “great excursion” appeared together as well.

Luxury Cruises Similarly to the case of commercial cruises, the scree plot for luxury cruises showed an inflection point at Concept 6. Consequently, by the elbow method, 6 concepts (vectors) were selected to continue the analysis. For each concept, the words (components) with more relative weight were selected from the table accompanying the scree plot. Singular values 35

Singular value % explained

0 0

Concept

The concepts described by the words that carried the higher weight (greatest absolute valued coefficients) were as follows: Concepts for Luxury Cruises: Concept 1: +great + food + service

Concept 2: + great + cabin / -bad-service Concept 3: +average + food / -bad-room Concept 4: +service + cabin + food / -ship Concept 5: + ship + cabin + food / -experience-good Concept 6: + excursion / -service-ship An inspection of sign distribution indicates that positive signs of the word coefficients in Concept 1 show co-occurrence of these word across the documents. Consumers from the sample of luxury cruises gave great importance to food and service. Concept 2, shows two groups, where “great” appears only with “cabin” and “bad” with “service” only. Concept 5 shows a group of opinions, where “ship”, “cabin” and “food” appear together. Conclusions Opinion mining was utilized to examine the needs differences of consumers who sail on commercial and luxury cruise lines. The goal of this study was to identify the key role of certain amenities on costumer choice across cruise lines. Opinion mining is a technique utilized by business organizations to effectively meet consumer needs. Findings from this study showed differences of consumer’s choice criteria at the time of booking commercial and luxury cruise lines. One hand, consumers traveling in either commercial or luxury cruises found great service, great food, and great excursions as key factors for cruise selection. On the other hand, this study found that those costumers that chose to travel in luxury cruises were more vocal at the time of expressing negative opinion when their needs regarding quality of service and room conditions did not meet their needs. This result reveals that those who chose to travel in luxury cruises have higher expectations and tend to be more critical about their experiences onboard. Acknowledgements The authors of this study are grateful to the students John Lizano, David Fernandez, Sage Rains, Rebeca Abreu, Nicolas Sousa, Rafael Gomez, Daylan Fuentes, Claudia Perez, Diego Martinez, Dianelis Lopez, Adriano Fernandez, Veronica Benitez, Patricia Milanes for their participation and valuable contribution in the study. The authors are thankful to Dr. Jorge R. Gonzalez for his support in the completion of the first stage of this study. The authors are also thankful to the editor-in-chief and reviewer for providing constructive feedback which improved the quality of the paper.

References Han, J., Kamber, M., Pei, J. (2012). Data mining: concepts and techniques. Morgan & Kaufmann. Elsevier. Han, H., Mankad, S., Gavirneni, N., & Verma, R. (2016). What guests really think of your hotel: text analytics of online customer reviews. Cornell Hospitality Report, 16(2), 3-17.