Leveraging Machine Learning to Optimize Office Hour Queues in Computer Science Courses; Bridging the Gap in Reluctant Evacuations by Looking at Its Factors Through the Eyes of Neuroscientists and Psychologists

Natural disasters, particularly water-related phenomena like hurricanes and tsunamis, have become increasingly frequent and severe due to climate change. As a result, the challenge of effectively responding to and mitigating their devastating effects becomes more urgent. This inevitability prompted my exploration of understanding the psychological and neurological factors that influence human decision-making during such crises. Although I am not someone who is immediately affected by the unfortunate consequence of natural disasters, it is something I thought of looking into due to the severity of the issue and difficulty of improving. I saw it as a way of challenging myself to utilize what I learned in STS 4500 and 4600 to view a problem in a way I wouldn’t have before. The technical project I worked on is unrelated and focused on improving an office hour queue system utilized in some computer science courses at the University of Virginia by implementing forms of machine learning to identify trends in student data and experience collected at the end of every semester and implement solutions based on trends found. These two projects aren’t connected, however, they highlighted to me the importance of integrating both technical and sociotechnical viewpoints into designing a system and broadened the scopes of a sociotechnical system to consider ethical and moral implications.
My technical study involved looking into improving the office hour queue system used in some computer science courses here at the University of Virginia. As a computer science student, I felt dissatisfied with the way office hours were done in some classes. I spent too much time in queue waiting for assistance to receive minimal aid or didn’t receive any due to time running out. Thus, after some research, I found a professor that was starting a new research topic regarding trying to improve office hours by utilizing machine learning concepts and algorithms. Initially, the research was limited to learning machine learning concepts, such as data cleaning and supervised learning, whilst data was being collected, as data was only accessible at the end of each semester. As time passed and more data was collected, a new implementation of a group queue was tested and data was observed to find trends and challenges with the existing system. Overall, a significant finding was that students don’t utilize office hours as efficiently as they should. In other words, students relied heavily on office hours right before difficult homework assignments were due, causing major traffic, inefficient aid, unhappy students, and stressed out professors and teaching assistants.
In my STS research, I focused on looking into the psychological and neurological factors that influence individual and community responses to natural disasters, particularly the decision to evacuate. My research revealed an overlooked insight: the "freeze" response, a neurological reaction to acute stress that can lead to decision paralysis in the face of danger. This response, in conjunction with cognitive biases such as risk misperception and emotional attachment to belongings, helped explain why many people hesitate to evacuate even when faced with clear threats. In addition to these factors, I found that distrust between communities and public authorities also plays a significant role in shaping evacuation behavior. Communities that have experienced previous failures in disaster response are more likely to resist evacuation orders due to a lack of trust in the efficacy of the authorities’ guidance while communities that are tight-knit with their authorities tend to listen to warnings and successfully evacuate. Understanding these psychological and social dynamics is critical in framing evacuation messaging that resonates with at-risk populations and fosters compliance with evacuation orders.
Overall, both projects were challenges that I took advantage of to test and gain knowledge applicable to sociotechnical systems. The technical research study I joined put me in a role where I had power to make a difference within a group by testing and implementing changes derived from a team. Working on the STS research paper and prospectus broadened my perspective of sociotechnical systems and all the ethics and morals that must be considered when making decisions, aligning my own morals with the ideas of engineering practice to prepare me for when I graduate. Although two different projects, they both served a purpose of bettering my understanding of sociotechnical systems as a whole by breaking them down and providing insights as to how each part plays its role within a system.