Exploring Gun Violence in Chicago
65,000+ records, three regression models, one 58-page report.
0+
Data points
0+
Lines of code
0 pages
Report length
Data Cleaning
Pandas, NumPy, and DuckDB process 65,000+ datapoints into an analysis-ready dataset.
Exploratory Data Analysis
Folium (built on Leaflet.js) maps whether certain Chicago zip codes experience higher rates of gunshot violence.
Subquestion I: Race, age, zip code & education
A Negative Binomial Generalized Linear Model tests the interplay between race, age, zip code, educational attainment, and likelihood of experiencing gun violence. The model is statistically significant: lower ages, specific zip codes, and specific races significantly increase one's chances.
Subquestion II: Seasonal variation
One-hot encoded month indicators show most months are significant in explaining gunshot violence; 78% of the variation is explained by the model, with an MSE of 0.001. July has the highest occurrence.
Subquestion III: Test scores & violence
Ranking Chicago zip codes by average standardized test scores yields a significant model test statistic of 10.60: one rank higher (i.e. lower average test scores) relative to other zip codes is associated with 1.72 higher gunshot injuries.