What's shown
Traffic deaths and serious injuries on surface streets across the City of Houston, 2016–2025 plus a partial 2026. Source: TxDOT CRIS public crash extracts. The boundary is the City's full-purpose (full-service) service area; limited-purpose annexation areas are not shown.
Which roads
This dashboard excludes limited-access freeways, tollways, ramps and frontage roads (a different facility type, and their crashes snap onto nearby streets), but keeps at-grade arterials regardless of owner, including state-owned ones like S Main / US-90A, SH 6, and Westheimer / FM 1093. This matches Austin's Vision Zero dashboard and the City's own High Injury Network, which include state arterials.
Road ownership
Each street and crash is labeled city-owned vs TxDOT-owned (from TxDOT's roadway inventory) so the dashboard can show the state-owned share; the City must partner with TxDOT to redesign those. A street is counted as TxDOT-owned only where it runs parallel to a state roadway, so a city street that merely bridges over a freeway is not mislabeled.
How crashes map to streets
Each crash is assigned to its nearest street within 200 ft (median 4 ft).
Definitions
Killed or seriously injured (KSI) = the K + A levels on the KABCO injury scale.
Years of life lost = years before age 75 summed over those killed (CDC YPLL convention). The public extract records a victim age for only about half of fatalities, so the rest are estimated at the average, making the total an estimate.
Neighborhood income = the median household income of the Census block group around each crash (an area characteristic, not the individual's).
Counts, not rates
Streets are ranked and shaded by the number of people killed or seriously injured, not a rate per traffic volume. This is the Vision Zero convention (and how the City's own High Injury Network is built): the goal is zero deaths, so attention goes where the most harm actually occurs. Counts are not normalized for exposure (vehicle miles traveled or foot traffic), which is unavailable for most local streets. A crash rate is also unstable on quiet streets, where a single crash produces a large per-mile-driven rate. On the busier streets where traffic counts exist, the harm remains about as concentrated after dividing by vehicle-miles traveled, which indicates specific dangerous roads rather than heavy traffic alone; separating road-design risk from exposure is a task for the systemic risk model (in development).
Crash costs
Each person hurt in a crash is assigned the FHWA per-person cost for their injury severity on the KABCO scale (FHWA-SA-25-021, 2024 dollars), summed across everyone in the crashes in view. Economic cost is the societal economic loss: medical care, lost productivity, property damage, and emergency, legal, and insurance services. It is a loss spread across victims, insurers, employers, and government, not a single public budget line.
The comprehensive figure adds the monetized value of lost quality and length of life. Its fatal portion values the same deaths the years-of-life-lost figure already counts (in years rather than dollars), so the two are different lenses on one loss and should not be added together.
Costs are summed per person, so a crash that kills or injures several people counts each of them. Injuries with an unrecorded severity carry no cost, and reported crashes undercount the true toll, so every figure shown is a conservative floor. Costs are national figures; no state adjustment is applied.
Data sources
Every dataset here is public. Each link goes to the official source.
Crashes · TxDOT Crash Records Information System (CRIS), public extracts, 2016–2025 plus partial 2026. txdot.gov · cris.txdot.gov
Street network & sidewalks · OpenStreetMap. openstreetmap.org
Road ownership (city vs TxDOT) · TxDOT Roadway Inventory. TxDOT Open Data Portal
Street design (speed, lanes, width, median) & traffic volume · City of Houston Public Works. COH GIS Open Data
High Injury Network overlay · City of Houston Vision Zero, official 2022 HIN. houstontx.gov/visionzero
Proactive Safety Network overlay · this project's research model (a concept network scored on street design, not crash history; not an official City designation). github.com/WrenVin/PharisFellowshipVisionZero
Neighborhood income · U.S. Census American Community Survey, 2023 5-year estimates (block group). census.gov/acs
Crash costs · FHWA, Updated Crash Costs for Highway Safety Analysis (FHWA-SA-25-021), KABCO per-person costs in 2024 dollars. highways.dot.gov
City boundary, council districts, super neighborhoods · City of Houston GIS. COH GIS Open Data
The specific source layers, vintages, and join methods are documented in the project CODEBOOK on GitHub.