Mapping Philadelphia's housing crisis, connectivity blind spots, and AI intervention traces with one public-facing civic intelligence platform.
The Survey · Live Numbers
Real addresses. Real violations. Pulled from six OpenDataPhilly sources and scored by machine intelligence - updated automatically.
Blighted parcels tracked across all Philadelphia zip codes
Active L&I violations on record, updated daily
All 159 mapped with real-time blight risk scores
Machine learning blight index per property
The Intelligence Layer
Every parcel in Philadelphia gets a machine-calculated risk index from 0-100. Here's how we build it.
Active L&I violations - structural, electrical, sanitation - weighted by severity and recency.
Confirmed vacant parcels from the Philadelphia Vacant Property database and utility shutoffs.
Years of unpaid property taxes compound the risk score exponentially past year two.
Eviction activity signals instability. High eviction density lifts scores for the whole block.
Surrounding parcel conditions and historical blight density factor in via spatial clustering.
All signals are passed through a gradient-boosted model trained on Philadelphia housing outcomes.
The Name & Its Meaning
In 1683, Thomas Holme laid a grid across a swampy river peninsula and called it Philadelphia. As William Penn's Surveyor General, he didn't draw mere lines on parchment - he mapped a future, ensuring the city would grow with intentionality.
Three hundred and forty-two years later, that grid still exists - but stretching across it are over 21,000 vacant properties, 4,000+ code violations, and thousands of eviction filings. The city Thomas Holme imagined is in crisis.
The Holmes Project carries his name as both tribute and mission. Every vacant building, every blight score, every eviction filing is a data point in a new survey of the city we inherited.
"Building AI for Philly's Future"
Philly Codefest 2026 · Submission theme
The Platform
Explore vacant properties, neighborhood risk, and the new connectivity layer in one map-first civic interface built for judges, residents, and field teams.
Ask for neighborhood summaries, policy briefs, tract risk explanations, and property-level context with Cloudflare Workers AI streaming through the platform.
Holmes now ingests housing data, Philadelphia connectivity tracts, public Wi-Fi sites, and Glass Box audit logs so every major section runs on real inputs.
Challenge
A live Leaflet connectivity map that scores tract-level risk, surfaces equity patterns, and explains local public Wi-Fi fallback in plain English.
Challenge
An audit dashboard that parses shared inhibitor logs into timelines, action summaries, and policy-trigger views for non-technical review.
Challenge
A platform that strengthens shared culture and fosters meaningful community connection by making civic data accessible, reflecting diverse experiences, and creating environments where collaboration and collective growth can thrive.