Micro-mobility blackspots | Paris 2024

What this map shows

Police-recorded injury accidents in Paris in 2024 involving at least one bicycle, powered personal mobility device or pedestrian. It maps 2,248 unique accidents out of 2,251 eligible accidents; three with invalid coordinates are omitted. These are accident counts, not counts of casualties or individual road users.

Full-year mapped totals: 1,196 bicycle-involving, 311 powered-EDP-involving and 1,046 pedestrian-involving accidents. Categories overlap: an accident can involve several included user categories, but is counted only once in each filtered hexagon. The full-year map contains 1,803 occupied cells, with a maximum of 6 accidents per cell.

Geometry and colours

Regular pointy-topped hexagons are 50 metres flat-to-flat (side length 50/√3 metres), not 50 × 50 metre squares. The fixed grid uses the supplied integer axial q/r coordinates in Lambert-93, EPSG:2154, relative to Paris centre (2.3522° E, 48.8566° N). The map uses equal metric x/y scales, with north upwards; no spherical re-projection is applied.

The discrete absolute scale is fixed at 1, 2, 3, 4, 5, 6+ accidents across all filters. Low-frequency cells are lavender; the highest frequencies are deep purple/near-black, retained against the dark basemap by pale-violet outlines. Counts above six, if future data are substituted, saturate at 6+.

Controls

• All day: 00:00–24:00 local time.
• AM: 07:00–09:00, 186 accidents across all included users.
• PM: 17:00–19:00, 316 accidents.
• Rush: the noncontiguous union of AM and PM, 502 accidents. The hours between the two peaks are not included.
• Drag the histogram brush or either handle for one custom continuous local-time interval. Interval starts are inclusive; ends are exclusive. Histogram bars use 30-minute bins, while accident filtering uses the supplied minute values.
• User controls select all eligible accidents, bicycle-involving, powered-EDP-involving or pedestrian-involving accidents. The histogram background reflects the selected user group; purple bars show the time-filtered selection.
• Scroll or use +/− to zoom; drag the map to pan. The metric 50 m grid never changes with zoom. FIT restores the map extent; RESET restores all users, all day and the map extent.
• Hover cells for unique-accident frequency, overlapping user-category counts and local accident times.

Definitions and limitations

BAAC bicycle vehicle categories 1 and 80 include electrically assisted bicycles. Powered EDP is vehicle category 50: it includes e-scooters and other powered personal mobility devices, which cannot be separately identified in this extract. Pedestrians are user category catu 3. Motor scooters, cars and nonpowered EDP are not included as target user groups; they may still be other parties in an eligible accident.

This is a descriptive frequency map, not exposure-adjusted risk, an insurance/pricing score, or evidence of validated intersection-level danger. Counts are not adjusted for journeys, traffic volumes, distances or reporting completeness. BAAC coordinates can be imprecise; 50 m aggregation and zooming do not improve their original accuracy. No precise intersection is certified as a blackspot. Casualty fields, if present in the source CSV, describe the whole accident and are not used to weight cell frequencies.

The street and arrondissement basemap is official Ville de Paris contextual data available when prepared, not a historical reconstruction of the 2024 street network. It is inlined and does not require network requests when the plot renders.

Sources

• BAAC annual road injury accident databases, through 2024
• Ville de Paris — Voies
• Ville de Paris — Arrondissements

The existing accident CSV and inlined official basemap were preserved; only rendering, controls and explanatory metadata were refined.