PA 470
Machine Learning in the Public Sector
PA 470 - Artificial Intelligence and Machine Learning in the Public Sector is a course taught by Eric Langowski as a part of the Civic Analytics curriculum at UIC. In the course, we worked through building an machine learning model to predict the sale prices of Cook county single family homes.
For my final project, I decided to do a replication study of Estimated Childhood Lead Exposure From Drinking Water in Chicago (Huynh et al., 2024). In the process, I learned a ton about Chicago’s lead contaminated water issues (Did you know that lead service lines were MANDATED by the city of Chicago until 1986 and as a result Chicago has 400,000 lead service lines).
The modeling involved a ton of data cleaning to aggregate data from the City of Chicago Department of Water Management Lead Test data, US Decennial Census, American Community Survey, Chicago Data Portal, and Chicago Health Atlas. Unfortunately, predicting lead exposure at the block level for a city as big as Chicago is quite a big problem. In the end I limited my models to elastic net logistic regression and random forest and limited my area of interest to the West Side of Chicago. But it was definitely a cool project, and you can check out my work at the above website.