Карта ДТП v1.0. 👉 База знаний о проекте: https://github.com/dtpstat/dtp-project/wiki
-
Updated
Jul 13, 2025 - Python
Карта ДТП v1.0. 👉 База знаний о проекте: https://github.com/dtpstat/dtp-project/wiki
The Safer Streets Priority Finder enables you to analyze the risk to bicyclists, pedestrians, and motor vehicles on your community’s roads.
roadway.report is a map of every traffic death in the USA
Бекенд и парсер. 👉 База знаний о проекте: https://github.com/dtpstat/dtp-project/wiki
Vision Zero Report Card — 7 years of NHTSA FARS truck-involved fatalities across 19 US cities. Reproducible Python pipeline + filtered dataset under CC BY 4.0.
Suri watches a street crossing and measures its close calls: SAM 3.1 tracks pedestrians and cars, code computes post-encroachment time, and a Muse Wonder agent verifies each near miss with tools while Muse Spark explains the blurred clip. Output: a Spanish inspection report for traffic engineers.
Download, filter, and aggregate NHTSA FARS fatal-crash data with zero non-stdlib dependencies. Companion to the Vision Zero Report Card.
Geospatial AI for Vision Zero / Toward Zero Deaths — H3 crash hotspot detection, CARTO spatial analytics, and Gemini natural-language queries for state DOT safety programs (Hawaii DOT, Minnesota TZD)
An autonomous agent that reads San Francisco's street data every morning, decides on its own whether anything changed enough to act on, and publishes the mornings it declines along with the reasoning.
Twenty years of Toronto's killed-or-seriously-injured collisions, analysed at the correct grain. The published file is one row per person, not per collision — counting rows overstates every headline figure by 2.3x to 3.7x.
Can Quebec's published collision reports predict how badly a crash turns out? They carry real but small information, almost all of it about who was involved rather than the road. Four independent arguments establish the limit. Python package, CLI, interactive explorer, Quarto book.
Video agent built on VAST AI OS + NVIDIA Cosmos + YOLO + W&B Inference (VAST Builders Challenge)
Bike crash risk prediction for Boston's Hyde Park to Forest Hills corridor. GNN and logistic regression on MassDOT traffic/LTS data and Vision Zero crash records. Live map updated monthly.
To associate your repository with the vision-zero topic, visit your repo's landing page and select "manage topics."