In this course, students will practice interfacing with local clients to better understanding their data-related needs. This may involve helping clients collect, analyze, visualize, or model data in a meaningful way. Students will then present their findings to a larger audience at the end of the semester.
My preference is that students work on projects that can make a positive impact on the local Vermont community. There are no specific priority areas, but my preference is for my students to assist smaller companies and nonprofits with data-related endeavors that they otherwise could not undertake due to time or monetary constraints.
As a PhD Statistician, I would oversee each of the student collaborations with clients. I would ensure any analysis, visualization, modeling, etc. done by students is sensical, and I could warn clients of any potential limitations of the students’ work.
The students in this class would be well-trained quantitative analysts and data scientists. They would have the ability to collect data (e.g., via surveys or experimentation), visualize the data (e.g., via interactive graphics, web apps, tables/charts), model the data (e.g., via machine learning, regression), and coherently convey the results of their work to meet the needs and interests of their clients. While each project likely calls for a unique approach, students would have the tools to be flexible and adaptive to the given situation.
The course will be offered during our Spring 2022 semester and lasts for 12 weeks. While there are certainly opportunities for students to extend the collaboration with the client(s) beyond the semester, I intend for the entire community-based project to be completed within the semester.
I hope community partners understand that our students will do incredible work. That being said, they are students (not paid consultants), and despite being overseen by me, there is always the potential for students to produce work that doesn’t sufficiently meet the client’s needs. It’s also entirely possible that the questions of interest from the client are impossible to answer given the data they have (or haven’t) collected, so flexibility may be warranted.
Build a web app to interactively visualize mental health trends in Vermont by looking at inpatient/outpatient hospital records
Analyze a large-scale survey given to youth at a local after-school program, providing insight into after-school activities that were engaging to students and those that weren’t (and then help identify problematic survey questions and reform them as necessary)
Take a messy data set looking at timestamps of when students enter each of the three dining halls at Middlebury, clean the data, visualize and analyze trends in dining hall usage, and then provide suggestions about an opening/closing schedule that better meets students’ needs.