We will look at crime rates for each county in California for several years (can expand the geography and time range if we decide). We will analyze which factors are predictive of high/low crime rates. We will consider socio-economic indicators, demographic data and government expenditure as features.
Here are some datasets that can be useful:
1) Violent crime rate by county in California from year 2000 to 2013: https://catalog.data.gov/dataset/violent-crime-rate-9a68e
2) Income levels by county: https://data.census.gov/table/ACSST5Y2023.S1901?q=income%20by%20county&g=040XX00US06
3) Detailed government expenditure by county and category for California from Years 2002 to 2015 https://bythenumbers.sco.ca.gov/Raw-Data/Counties-Raw-Data-for-Fiscal-Years-2002-03-to-2015/esdm-5xr2/about_data
We will find more datasets and combine them, lots of socio-economic data is available by county. Our goal is to figure out which factors are predictive of crime rates, it is harder to conclude causality (for example higher police budget could cause lower crime rate, or high crime rate could cause higher police budget).