Conference Session

DSA Technical Session 3

Collection

2024 ASEE Annual Conference & Exposition

Authors

Aidan Kenny, Northeastern University; Andrew L Gillen, Northeastern University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

helps to communicate key findings [3]. By considering various datavisualizations methods as well as the design principles used to present them, understanding andinterpretation by the user can be improved.In the 1890’s, W.E.B. DuBois published artistic visualizations of African American civil rightsviolations that grabbed the public’s attention and clearly displays the intended conclusions [4].These pioneering data visualizations demonstrate the profound significance of creativity withinthe field of data representation and analysis. Through his use of intricate hand-drawn charts,graphs, and maps, DuBois demonstrated that the presentation of data could go beyond statisticsand become a powerful tool for conveying complex realities. For example

Conference Session

DSA Technical Session 5

Collection

2024 ASEE Annual Conference & Exposition

Authors

Nicolas Leger, Florida International University; Maimuna Begum Kali, Florida International University; Stephanie Jill Lunn, Florida International University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

. For instance, inthe realm of social media, data science has brought about a paradigm shift in the understandingof communication. It has moved beyond analyzing communication as signs or discourse and nowencompasses the collection, storage, and processing of communication data. This expansion inperspective has opened up new possibilities for studying and leveraging social media platformsin various domains. For example, at the earlier stage of social media, Langlois et al. proposed anontological shift, suggesting that with the help of data science, “we must expand from the studyof communication as signs or discourse to include the study of communication as data collection,storage, and processing [5, p. 2].” Consequently, these new technologies

Conference Session

DSA Technical Session 8

Collection

2024 ASEE Annual Conference & Exposition

Authors

Neha Kardam, University of Washington; Denise Wilson, University of Washington; Sep Makhsous, University of Washington

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

://essay.utwente.nl/82090/[16] V. S. Sadanand, K. R. R. Guruvyas, P. P. Patil, J. Janardhan Acharya, and S. Gunakimath Suryakanth, “An automated essay evaluation system using natural language processing and sentiment analysi,” International Journal of Electrical and Computer Engineering (IJECE), vol. 12, no. 6, p. 6585, Dec. 2022, doi: 10.11591/ijece.v12i6.pp6585-6593.[17] F. Dalipi, K. Zdravkova, and F. Ahlgren, “Sentiment Analysis of Students’ Feedback in MOOCs: A Systematic Literature Review,” Frontiers in Artificial Intelligence, vol. 4, Sep. 2021, doi: 10.3389/frai.2021.728708.[18] E. Mayfield, M. Madaio, S. Prabhumoye, D. Gerritsen, B. McLaughlin, E. Dixon-Román, and A. W. Black, "Equity beyond bias in language

Conference Session

DSA Technical Session 8

Collection

2024 ASEE Annual Conference & Exposition

Authors

Neha Kardam, University of Washington; Denise Wilson, University of Washington

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

tocontribute to the development of educational research methodologies. It emphasizes the potentialcollaboration between automated coding systems and human expertise in interpreting studentfeedback data.Literature ReviewOver 16 million people are enrolled as undergraduates in colleges and universities in the US [6].Understanding the lived experiences of these students on a broad scale including their satisfactionwith their education, learning outcomes, and intentions to persist in their careers requireseducation-based research that extends beyond the standard Likert-scale questions on surveys andstudent evaluations of teaching [1]. Augmenting surveys with short answer questions allowsresearchers and instructors to more effectively and more thoroughly

Conference Session

DSA Technical Session 1

Collection

2024 ASEE Annual Conference & Exposition

Authors

Ahmad Slim, The University of Arizona; Gregory L. Heileman, The University of Arizona; Melika Akbarsharifi, The University of Arizona; Kristina A Manasil, The University of Arizona; Ameer Slim, University of New Mexico

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

, therefore,is not merely an academic concern but a pivotal factor that could shape students’ educational tra-jectory and success. However, research in this domain often needs to be revised, particularly inunraveling the intricate causal relationships within the educational ecosystem. While numerousstudies have explored how various aspects of curriculum design affect student outcomes, they pre-dominantly employ conventional data analysis and machine learning methods. These approaches,while valuable, often lead to results that are correlative rather than causative. Therefore, the chal-lenge lies in moving beyond identifying patterns and correlations to understanding the underlyingcausal mechanisms. This gap in existing research is particularly evident

Conference Session

DSA Technical Session 5

Collection

2024 ASEE Annual Conference & Exposition

Authors

Mehmet Ergezer, Wentworth Institute of Technology; Mark Mixer, Wentworth Institute of Technology; Weijie Pang, Wentworth Institute of Technology

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

] S. Negash, “Business intelligence,” Communications of the association for information systems, vol. 13, no. 1, p. 15, 2004. [5] S. Siuly and Y. Zhang, “Medical big data: neurological diseases diagnosis through medical data analysis,” Data Science and Engineering, vol. 1, pp. 54–64, 2016. [6] D. A. Jenkins, M. Sperrin, G. P. Martin, and N. Peek, “Dynamic models to predict health outcomes: current status and methodological challenges,” Diagnostic and prognostic research, vol. 2, no. 1, pp. 1–9, 2018. [7] J. Chen, K. Li, H. Rong, K. Bilal, N. Yang, and K. Li, “A disease diagnosis and treatment recommendation system based on big data mining and cloud computing,” Information Sciences, vol. 435, pp. 124–149, 2018. [8] L. Sun, C. Liu

Conference Session

DSA Technical Session 8

Collection

2024 ASEE Annual Conference & Exposition

Authors

Paula Francisca Larrondo, Queen's University; Brian M Frank P.Eng., Queen's University; Julian Ortiz, Queen's University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

, graders, or course instructors teachingcourses focused on complex engineering problem-solving. We detail how fine-tuning an LLMwith a small dataset from diverse problem scenarios achieves classification accuracies close toapproximately 80%, even in new problems not included in the fine-tuning process. Traditionally,open-source LLMs, like BERT, have been fine-tuned in large datasets for specific domain tasks.Our results suggest this may not be as critical in achieving good performances as previouslythought. Our findings demonstrated the potential for applying AI-supported personalizedfeedback through high-level prompts incentivizing students to critically self-assess theirproblem-solving process and communication. However, this study also

Conference Session

DSA Technical Session 6

Collection

2024 ASEE Annual Conference & Exposition

Authors

Marjan Eggermont, University of Calgary

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

, Institute of Electrical andElectronics Engineers, 2013, 19 (12), pp. 2396 - 2405.[2] Jansen, Y. Dragicevic, P. Isenberg, P. Alexander, J. Karnik, A. Kildal, J. Subramanian, S. andHornbæk, K. Opportunities and Challenges for Data Physicalization. CHI 2015, April 18–23,2015, Seoul, Republic of Korea.[3] Data Physicalization Gallery, http://dataphys.org/list/gallery/[4] Noë, Alva. “What Art Unveils.” The New York Times, The New York Times, 5 Oct. 2015,archive.nytimes.com/opinionator.blogs.nytimes.com/2015/10/05/what-art-unveils/.[5] Rohit Ashok Khot, Larissa Hjorth, and Florian ‘Floyd’ Mueller. Understanding PhysicalActivity through 3D Printed Material Artifacts. CHI 2014, April 26 - May 01, 2014, Toronto,ON, Canada.[6] Vetria L. Byrd. Usability of

Conference Session

DSA Technical Session 4

Collection

2024 ASEE Annual Conference & Exposition

Authors

Fengbo Ma, Northeastern University; Xuemin Jin, Northeastern University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

an utterance, providing a finer time granularity. It captures the dynamic nature of speechemotions, going beyond the scope of utterance-level classification by capturing discreteemotional changes over time. The two different methods are illustrated in Figure 1. Figure 1: Speech Emotion Recognition (SER) methods2.1 Mel-Frequency Cepstral Coefficient (MFCC)Mel-frequency cepstral coefficients is a widely used feature extraction technique in the field ofaudio signal processing and speech recognition [3]. It was first proposed by S.B. Davis, and P.Mermelstein [4] in 1980. MFCC is crafted based on the auditory perception of humans, whichtypically does not register frequencies above 1 kHz.Essentially, the MFCC framework is

Conference Session

DSA Technical Session 2

Collection

2024 ASEE Annual Conference & Exposition

Authors

Ben D Radhakrishnan, National University; James Jay Jaurez, National University; Nelson Altamirano, National University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

companies, lack ofadditional rebates and tax credits. One other key element could be that the public might needmore solar energy generation (and use) awareness through effective public education. It couldalso be financial (upfront cost and lack of financing resources).US Solar Energy Forecasting (2023 – 2030) and State-Level K-MeansClustering Data AnalysisThe current solar generation trend in the US States was discussed with charts. US Solargeneration data was available for 33 years and a significant increase in solar generation trendwas noticed starting in 2018 in 46 States (the following states had zero: AK, ND, NH, WV, andDC). This section will look at the future of solar generation based on significant data from thepast and analyze the State

Conference Session

DSA Technical Session 1

Collection

2024 ASEE Annual Conference & Exposition

Authors

Betul Bilgin, The University of Illinois at Chicago; Naomi Groza, The University of Illinois at Chicago

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

engineering is driving innovation, increasingmanufacturing efficiency, and fostering collaboration with other industries [6]. It is also criticalto form a team proficient in skills, a comprehensive understanding of data science, and aconcentrated emphasis on the genuine requirements pivotal to the digital transformation withinthe chemical industry [7], [8].The integration of data science into chemical engineering curricula is crucial for preparingstudents to meet modern industry challenges [1]. This integration can be achieved through themodification of existing courses, introduction of electives, or the creation of specializedprograms [9]. Data science principles, including data management, statistical and machinelearning, and visualization, are

Conference Session

DSA Technical Session 3

Collection

2024 ASEE Annual Conference & Exposition

Authors

Tushar Ojha, University of New Mexico; Don Hush, University of New Mexico

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

graduate. Numerous other studies have painted asimilar picture of excess credit accumulation [11, 23, 14, 15].This has attracted attention from university and government administrators alike because credithours presumably have direct relevance for student success and finances as well as for publicfinances. Excess credits are accompanied by an increase in time–to-degree [16] [1] [8], and abachelor’s-degree-seeking student can expect to pay $68,153 in cost of attendance and lostwages [7] for an additional year of college. As for the public cost of extra credits, if all bachelor’sstudents pursued just three extra credit hours beyond their degree requirements, it would result inan annual expense of $1.5 billion for Americans [7]. A 2004 study by the

Conference Session

DSA Technical Session 6

Collection

2024 ASEE Annual Conference & Exposition

Authors

Smitesh Bakrania, Rowan University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA)

learning management systems (LMS) used during this time period 1collected vast amounts of data on individual students. While visual and verbal feedback that aclassroom environment offered were lost in the switch to online, data quantifying studentbehavior on the LMS was readily available even for small class sizes. Through analysis,educators for these smaller classes have the potential to utilize this information to get a bettersense of student needs and adapt coursework accordingly, similar to how MOOCs use thisinformation more broadly.Learning analytics (LA) is a growing field in educational research. LA is used by educators tomake critical decisions about the educational experience for students

Conference Session

DSA Technical Session 7

Collection

2024 ASEE Annual Conference & Exposition

Authors

Abdulrahman Alsharif, Virginia Polytechnic Institute and State University; Andrew Katz, Virginia Polytechnic Institute and State University

Tagged Topics

Data Science & Analytics Constituent Committee (DSA), Diversity

, while ensuring data privacy through local deployments. Futurework could explore more advanced GAI models to further streamline the clustering and codegeneration workflow.Introduction With the rise of generative AI (GAI) tools such as GPT-4, Claude, Llama, and others,interpreting textual data and generating coherent responses is now possible with powerfulcapabilities. Traditionally, analyzing large qualitative datasets like student survey responsesrequires extensive time. Natural language processing (NLP) can help decrease the time neededand provide a solution to this obstacle. However, NLP cannot perform equally well on everytask; some tasks see better performance than others. Because each model is trained on differentdatasets, choosing