Hume network-wide AADT — measured where counted, modelled elsewhere
Annual daily traffic for every road section in Hume City. Arterials use measured DTP counts; local/collector roads are modelled (R² 0.874), anchored by 272 measured SCATS intersection volumes. Draft for review.
12,160road sections
1,186measured arterial (DTP)
272SCATS measured intersections
0.874model R² (local roads)
Coverage map
measured arterial modelled local/collector SCATS measured intersection (size = volume)
What moved AADT accuracy
Every map-derived POI/traffic proxy (Mapbox, Overture, buildings) was redundant with features the model already had. Only real measured SCATS counts added independent signal.
SCATS validates against measured AADT & distribution by tier
Method, in brief
| Tier | Source | Sections | Basis |
|---|---|---|---|
| Measured arterial | DTP Traffic Volume | 1,186 | Loop/tube counts on declared network |
| Measured intersection | DTP SCATS (May 2026) | 272 sites | 15-min loop counts, summed to typical weekday entering volume |
| Modelled | HistGB on OSM + land-use + SCATS features | 10,974 | R² 0.874, 5-fold CV |
Sources: DTP/VicRoads Traffic Volume & Traffic Signal Volume (SCATS) open data; OpenStreetMap; Overture Maps. Modelled values are estimates — red-pen freely.