Police Transparency
Dashboard

The City of Lake Forest's police department needed a public-facing tool to communicate where crime incidents had occurred in the community — a standard transparency need for municipal police departments. But the raw data they had to work with was a messy export from their records management system: inconsistent address formats, unstandardized incident categories, and no spatial component.

Before this project, publishing an update meant manually cleaning the spreadsheet, geocoding addresses one by one, sorting incidents into display-ready categories, and hand-placing the data into whatever publishing workflow was in use. Each update cycle consumed 60 to 90 minutes of staff time — a significant load for a recurring operational task.

Crime incident addresses can't simply be published at the parcel level. Specific addresses can identify victims, expose residents, or create legal exposure. Any public tool had to find the right level of geographic detail — informative enough to be useful, aggregated enough to protect individual addresses.

Built an ArcGIS Experience Builder application as the public-facing interface — a spatial dashboard giving residents a clear, map-based view of where incident types have occurred across the city. The Experience Builder platform made it straightforward to configure layout, filtering controls, and display logic without custom front-end development.

Wrote a Python script that takes the raw police department export and handles the full processing chain automatically: geocoding addresses from the messy PD format, categorizing and sorting incidents into display-ready types, anonymizing location points by rolling them up to 100-block zones, and posting the processed dataset directly to ArcGIS Online via the REST API. An update that previously required 60–90 minutes of manual work now runs in approximately 5 minutes.

The 100-block zone approach was the key design choice. Rather than suppress all location data or display precise addresses, the script rolls each incident's coordinates to the midpoint of its 100 block (e.g., "400–499 Main St" becomes a single representative point). This gives residents meaningful neighborhood-level context — "there have been vehicle break-ins in this part of town" — without exposing the precise residential or commercial address involved. It's the right level of aggregation for public consumption.

The update process dropped from 60–90 minutes of manual staff work to a 5-minute automated script run — approximately 92% time reduction. Staff no longer need to manually clean addresses, geocode individual records, or hand-sort categories before publishing. The pipeline handles all of it consistently, every time.

The City of Lake Forest now has a maintained, public-facing tool that communicates crime incident patterns to residents — meeting a genuine community transparency need that previously had no adequate solution.

The project demonstrates that privacy-sensitive public data doesn't require choosing between transparency and protection. The 100-block aggregation approach is a replicable pattern for any municipality facing the same tradeoff with crime data, code enforcement records, or similar incident datasets.