The Data
RideScore DC publishes two artifacts: a data package (Parquet files, the road network and safety scores) and a serving bundle (SQL, deciding what a map on the website may show). npm run data downloads both and loads them into your local database; the Full Stack Developer Guide covers the loading commands.
The currently published data set that is shown on the website comes from public Open Data DC: the roadway block network, crash records, and the Washington DC boundary.
The three datasets
| Dataset | Rows | One row is |
|---|---|---|
road_segment | 13,829 | one block of road, with attributes and geometry |
crashes | 2,224 | one crash that injured or killed a cyclist, positioned |
ridescore_v1_scores | 13,829 | the safety score and its components for one block |
road_segment
| Column | Meaning |
|---|---|
segment_id | the block's durable identity (DDOT BLOCKKEY); feedback is stored against it |
tile_id | integer for MapLibre feature-state — per build, not durable |
route_id, route_name | the street the block belongs to |
function | Local, Collector, Minor Arterial, Principal/Primary Arterial, Other Freeway and Expressway, Interstate, Other |
num_lanes_raw / num_lanes | lane count as published / with missing values filled in |
speed_limit_raw / speed_limit | mph as published / with missing values filled in |
bike_facility_type | none, painted_lane, buffered_lane, protected_track |
parking_presence, road_width, pavement_condition, slow_street | street attributes as published (width is the total cross-section) |
len | segment length |
crash_count_5yr, serious_injury_count_5yr, fatal_count_5yr | crashes attached to the block within the run's five-year window |
geometry | simplified line geometry |