Case Study · Policy + Data Analysis
Arrest Trends, Policy Reform, and Racial Disparities
A comparative analysis of how policing patterns shift over time, across cities, and across communities when you stop looking only at isolated moments.
This independent project examined public safety data across New York City, Washington, D.C., and Los Angeles to answer a harder question than “did reform work?” It asked what changed, when, for whom, and what remains structurally consistent underneath the headlines.
The problem
Public debate moves fast. Structural patterns do not.
Conversations about policing reform often rely on short time windows, anecdotal evidence, or one city at a time. That makes it easy to miss recurring seasonal spikes, policy lag effects, or disparities that remain stable even when total arrests fall. The project set out to build a longer, more comparative view that could support better policy interpretation.
My role
Research owner
I led the project from data cleaning and exploratory work through modeling, visualization, and written policy argument.
Translator
I converted technical findings into a narrative a policymaker, academic audience, or journalist could follow without flattening the nuance.
Methods + tools
Analytical arc
Prepare the data
Standardized offense categories, dates, and city labels while auditing missingness, outliers, and comparability issues.
Surface long-run patterns
Built monthly totals, rolling averages, and seasonal views to distinguish noise from durable movement.
Map reforms to the timeline
Aligned major policy changes and enforcement shifts with observed trend breaks rather than assuming a clean before-and-after story.
Test the claim
Used difference-in-differences style comparisons to evaluate whether reform periods coincided with meaningful shifts in outcomes.
Write the argument
Paired charts with explicit caveats so the narrative stayed persuasive without pretending the data answered more than it could.
What emerged
Seasonality matters
Summer arrest spikes appeared repeatedly, suggesting that short-window comparisons can misread normal cyclical changes as policy effects.
Reform effects are uneven
Some inflection points aligned with reform periods, but the distribution of impact was inconsistent across cities and racial groups.
Equity cannot be inferred from totals alone
Declines in aggregate arrests do not automatically mean disparities are shrinking; composition and burden still matter.
Artifacts
The strongest evidence here is the combination of reproducible analysis, interpretive writing, and a comparative frame that resists simplistic conclusions.
Next move
- Extend the analysis beyond 2024 and add more metropolitan areas for broader comparison.
- Bring in neighborhood-level indicators to better capture where disparities concentrate spatially.
- Translate the work into an interactive dashboard for policymakers, journalists, and public readers.