A Data Analysis Project

Detroit Crime,
Read Through the Weather

Nine years of police incidents meet daily temperature, rain and snow — to ask a simple question: how much does the weather outside shape the crime reported across the city?

791,583 incidents 3,441 days 2017-01-012026-06-03
+4.8%
crime per +10°F
citywide, all categories
+41%
hot vs. cold days
230 crimes/day average
+9.8%
Aggravated Assault
most heat-sensitive offense
23
offense categories
each profiled separately
The Report11 sections · 16 charts

The full analysis

A written, chart-by-chart walk through the whole story — from the headline temperature effect to the surprises hiding in the details.

  • How heat drives crime — and survives deseasonalizing
  • Which offenses ride the thermometer, which ignore it
  • Hot nights, hot weekends, and the geography of it all
  • Rain, snow, wind & heat waves — each isolated from temperature
Read the report
Interactive23 offenses + roll-ups

The weather explorer

Pick any single offense and watch the whole page reshape to its weather personality — live charts, plain-English summary, instant comparisons.

  • Temperature response, season and daily rhythm per category
  • Rain- and snow-sensitivity at a glance
  • Every category measured against the citywide baseline
  • Self-contained & instant — no loading, no internet
Open the explorer

How it was built

Detroit police incidents are aggregated to daily counts and joined to local daily weather. Temperature effects come from correlations and regressions; every precipitation, wind and heat-wave figure is reported holding temperature constant, so the weather signal is never just the season in disguise. Full methodology and caveats live at the end of the report.

Crime data: Detroit Open Data Portal Weather data: Open-Meteo historical archive Built with: Python · pandas · SciPy · Matplotlib