Tucson crime analysis

This project tested how reported crime in Tucson relates to neighborhood income and city streetlight locations. Our team combined four public datasets and used regression and classification models to examine two hypotheses.

Lower median household income was associated with higher crime counts in the data. Streetlight count was also positively associated with crime. That second result ran against our starting hypothesis. A likely explanation is that lights are installed in response to crime or alongside other features of dense areas. The data cannot establish cause.

What I worked on

The strongest result was methodological. A model can produce a clear number and still support only a narrow claim. The final report separates predictive fit from causal interpretation and records where the data is too weak to decide.