© OpenStreetMap contributors

Automatic offline bundle

Preparing data

Checking bundled accident data.

Average severity by ranked intersections

Smoothed Top-N average per state

Accidents per population

Filtered accident outcomes per 100,000 residents by state

Population-scaled outcome rates

Total accident rate compared with fatal and serious injury rates

Severity correlation

Severity % compared with severe crash rates per 100,000 residents by state

Average severity by ranked intersections

Smoothed Top-N average per region

Accidents per population

Filtered accident outcomes per 100,000 residents by region

Population-scaled outcome rates

Total accident rate compared with fatal and serious injury rates

Severity correlation

Severity % compared with severe crash rates per 100,000 residents by region

Data

Bundled normalized accident data loads automatically from the offline data bundle.

Accident data: Unfallatlas / Statistische Aemter des Bundes und der Laender.

Municipality data: Gemeindeverzeichnis / Destatis municipality directory extract, 2nd quarter 2026.

Reused under Datenlizenz Deutschland - Namensnennung - Version 2.0 / dl-de/by-2-0. Source data is processed, clustered, and analyzed by this app.

Project repository: kachurovskiy/sichereknoten.

Legal notice: Maxim Kachurovskiy, Konzeller Str. 26, 85635 Höhenkirchen-Siegertsbrunn, m.kachurovskiy@gmail.com.

Severity % metric

Analysis

Road-user focus

Select one or more to include only accidents involving at least one selected road user. Leave all unchecked to include all accidents.

About Severity %

What it measures

Severity % is a weighted share of severe outcomes at an inferred intersection. By default: (fatal + serious / 2) / total. Fatal accidents count once, serious-injury accidents count as half, and all accidents at the intersection form the denominator. In the raw formula, 100% severity means the weighted severe count equals the total accident count; with default weights, that means every known record at the intersection was fatal. The displayed value can also reach 100% when trend adjustment pushes a high raw score up to the metric cap.

Low accident totals are weighted conservatively, and the accident trend adjusts the result gradually once the yearly change is large enough to influence the score.

Why this focus

We do not have reliable traffic volume data for each intersection. Intersections with many recorded incidents are therefore not automatically the highest-severity locations; they may also be very highly loaded.

The metric focuses on intersections with higher recorded severity, using fatal and serious-injury outcomes to distinguish severe locations from high-volume ones.