SenseBox Network Sampling

Predicting cycling overtake exposure on street network edges from openSenseMap bicycle sensor data.

Idea

senseBox:bike devices record GPS tracks with measured lateral distances of overtaking vehicles. We match these measurements to OSM street network edges, enrich the edges with static traffic and infrastructure attributes, and test how well sparse sensor data can predict overtake exposure across the network — what hurts prediction more, starving the temporal dimension (collection frequency) or the spatial coverage (which edges get sampled)?

Setup

conda env create -f environment.yml
conda activate sensebox_network_sampling

Network preprocessing

senseboxbike_preprocessing.py builds the analysis network for Münster:

1. Acquisition — downloads the cyclable street network (network_type="bike": all ways cyclists may legally use, not just bike infrastructure) via OSMnx, topologically simplified, projected to EPSG:25832, cached as GraphML in input/.

2. Sidepath detection — Münster's cycle infrastructure is largely sidewalk-level tracks alongside roads. Car-free edges are flagged as sidepaths via the OSM tag is_sidepath=yes combined with a geometric check: within 12 m of a motor-traffic road unconditionally, within 12–30 m only if roughly parallel (bearing of longest segment within 25°).

3. Classification — each edge gets one riding-regime class (precedence-ordered, most specific first):

   class	meaning
   bicycle_street	Fahrradstraße, cars are guests
   roadside_track	separated track along a road (red tiles etc.)
   independent_path	cycleway/path away from motor traffic
   painted_lane	on-carriageway cycle lane
   pedestrian_zone	riding among pedestrians
   residential_street	low-speed residential streets
   service_way	access/parking/service ways
   offroad_track	unpaved field/forest tracks
   bus_lane	bus and bikes share the carriageway
   minor_road_shared	mixed traffic, minor connector roads
   main_road_shared	mixed traffic on an arterial road

   OSM tag booleans (has_lane_tag, is_cycling_street, is_sidepath, …) are kept as separate columns for modelling.

4. Audit & outputs — diagnostics (highway-tag crosstab, precedence audit, length-weighted class shares, small-multiple maps) are written to output/inspection/; the classified edges to input/muenster_edges_classified.gpkg; an overview map to output/network_overview.png.

Run with:

python network_preprocessing.py

The OSM download is cached — delete input/muenster_bike.graphml or call get_graph(force_download=True) to refresh.

Data

OpenStreetMap for the street network and static attributes, openSenseMap for the bicycle sensor measurements. Raw data lives in input/ and is not tracked in the repo.

Related work

Network semantics informed by the IIP bikeability pipeline (OSMBicycleInfrastructure, IP-OSeM-Backend), reimplemented in Python on a complete cyclable network.

Work in progress.