Police Response Deserts

Display Metric

Click any census tract to see full details

Neighborhood Lookup

Find your neighborhood's response time stats

Legend

Tract Details

Click a census tract on the map

Income vs Response Time

Key Findings

Mapping 911 Call Response Time Inequality Across Detroit Neighborhoods

Research Question

When you call 911, how quickly police arrive depends enormously on where you live. This project uses real Detroit 911 call-for-service records with timestamps to calculate actual median response times at the census tract level, maps the spatial inequality, and tests whether response time correlates with race and income after controlling for call volume and station proximity.

Methodology

Downloaded 297,417 Detroit 911 police calls from 2022 via ArcGIS REST API
Filtered to valid response times (0–240 minutes), retaining 281,725 calls
Spatially joined call locations to 627 Wayne County census tracts (TIGER/Line)
Aggregated by tract: median response, 90th percentile, day/night splits, call volume
Joined ACS 2022 demographics: % non-white (B03002), median income (B19013)
Computed distance from each tract centroid to nearest police precinct
Ran OLS multivariate regression to test equity predictors

Regression Results

OLS regression: dependent variable = tract median Priority 1 response time (n=270 tracts)

Variable	Coefficient	p-value	Result
Distance to nearest precinct (km)	+0.654	<0.001	★ Significant
Median household income	−0.147	0.048	★ Significant
% Non-white population	−0.142	0.068	Marginal (p<0.10)
Call volume per tract	+0.029	0.693	Not significant
Population density	−0.012	0.869	Not significant

R² = 0.257 · Model explains 26% of variance in tract-level response times

Key finding: After controlling for distance to the nearest precinct (the strongest operational predictor), lower-income tracts still receive statistically slower response (p=0.048). The race coefficient trends in the same direction but falls just short of conventional significance (p=0.068), consistent with partially income-mediated racial disparities in service access.

About This Project

Built as a portfolio data project demonstrating spatial analysis, equity research methodology, and full-stack GIS web development. All data is publicly available. Analysis code is available in the repository's notebooks/ folder.