Data science has played a transformative role across diverse fields in research, industry, and education in the 21st century. In addition to computational sciences such as computer science and electrical and computer engineering, other interdisciplinary fields such as civil and environmental engineering (CEE) have begun integrating data science into their curricula. Recently, the field has been disrupted by the emergence of powerful generative AI (GenAI) tools, such as OpenAI’s GPT-4, which powers the ChatGPT chatbot. These large language models can perform sophisticated tasks like interpreting large datasets, writing code for wrangling and analyzing data, brainstorming ideas, and explaining complex statistical and mathematical concepts. In CEE, the built environment generates extensive datasets, including air quality, structural health, and energy consumption, compelling future CEE practitioners to analyze and make sense of all aspects of the built environment. In CEE education, GenAI has the potential to make data science more accessible to students who may not have a strong background or interest in data analysis or coding. This paper addresses two primary research questions: (1) How do GenAI tools impact CEE students’ ability to interact with, clean, visualize, and interpret large datasets compared to the pre-GenAI era? (2) What are students' attitudes, perceptions, and experiences with GenAI tools for working with data in domain-specific CEE education? We answer these questions through the analysis of individual and team data analysis and visualization tasks and reflections produced by students across two offerings of an undergraduate CEE course, in which students are tasked with cleaning, visualizing, and analyzing a large air quality dataset collected using PurpleAir® sensors they deployed themselves on campus. The first iteration of the course was offered before GenAI tools were widely available in Spring 2023. In the second iteration, in Spring 2024, students were given the option of using a GenAI tool of their choice to help them complete their data processing and analysis coursework. Initial results show no significant difference in students’ grades between semesters. In Spring 2024, students’ choice of whether or not to use GenAI were guided by their initial perceptions of the tool. For instance, students who expressed critical views about the accuracy of GenAI output opted to not use the tool for their data tasks, whereas students who saw the potential of the tool to support automated data tasks chose to use the tool. Interestingly, most student’s use of GenAI went beyond data processing tasks and included asking the tool for advice or guidance, troubleshooting code, and web research. These findings also align with students’ reflections in which they self-reported the areas in which they learned the most and gained the most value. GenAI has the potential to transform CEE education, not only by making technical skills more accessible, but also by guiding and enhancing students’ critical thinking and problem solving processes. However, students’ cautious adoptions show a need for more structured education and support in order to increase the potential benefits to student learning and students’ trust in their use.
The full paper will be available to logged in and registered conference attendees once the conference starts on June 22, 2025, and to all visitors after the conference ends on June 25, 2025