The Pedestrian Network (pednet) was created by the DAV team at the City of Toronto, and it is based on the sidewalk inventory from Transportation Services, Toronto road centrelines, and manual collection from aerial imagery. Pednet is integrated with centerline intersections, traffic signals, pedestrian crosswalks and crossovers, traffic signal data from Transportation Services as well as other City of Toronto datasets.
Pednet was built using a variety of open source libraries such as NetworkX, Pandana, Quantum GIS, and Space syntax, as well as production mapping tools from ESRI’s ArcPro/ArcMap. The project source code can be found on DAV’s GitHub account here, which includes the semi-automated offsetting method from the Sidewalk Inventory and the analytical procedures undertaken.
Pednet is a data model resembling a network graph (edges and nodes) weighted by linear distance. Shortest routes were calculated from every building centroid in the city to the nearest nth amenity at the maximum distance of 5000m. Walk times were calculated in the nearest minutes, using the prescribed 1.0m/per-second velocity used by Transportation services. Two separate versions of pednet were created in this iteration of the project:
1) using actual linear distances as network weights, and
2) where crosswalks were “extended” by 20% of their length to impose additional impedance to their distances and walk times.
For every address within the City of Toronto, the walk times were calculated to various amenities like schools, libraries, hospitals, supermarkets, TTC stops and convenience stores see Section 3. Walk times were assigned to individual addresses as attributes. We then aggregated all these walk times to the census tract level and calculated the minimum, maximum, standard deviation, median, and average walk times. We used these aggregated values to both:
1) relate walkability measures to Statistics Canada Census data for socio-demographic analysis, as well as
2) the building footprints, pednet centerlines, and census tract area boundaries to be used in choropleth maps contained within the following sections.