Seattle - Automatization of updating curb situation in Seattle
Mapping projects, whether private, proprietary or public, have traditionally focused on automobile roads. Consequently, pedestrian-relevant information has often been overlooked or only mapped as attributes of the road. Even OpenStreetMap is missing a lot of sidewalk and other pedestrian information.
All pedestrians can use more information about the pedestrian environment before they approach it, but this is especially crucial for people with limited vision and mobility.
Municipalities need to collect this kind of information, but they store and disseminate it in a format very different from OpenStreetMap, because they look at sidewalk segments as ‘independent assets’, rather than a transportation network of its own.
accessmap.io currently runs off of data derived (via accessmap inference algorithms) from the municipal data to provide a trip planning service for people with limited mobility. The OpenSidewalks project seeks to clarify and define a holistic pedestrian data schema within OpenStreetMap, and develop tools to improve the flow of this information between municipalities and OpenStreetMap.
OpenSidewalks in Seattle
The OpenSidewalks project is converting the information from data.seattle.gov into its current schema and doing structured imports via mapathons, where citizen scientists pore over every sidewalk and intersection to add essential pedestrian information. The Seattle OpenStreetMap community has co-hosted these events and been amazingly supportive. As this information becomes more complete, accessmap.io will switch to OpenStreetMap as its source of pedestrian data, and feed back into OpenStreetMap by letting users report infrastructure as they see it.
Types of data
Data licenses: public domain with attribution, ODbL Compliance verified
Technical solutions
Forked the iD editor to add some presets based on OpenSidewalks
Using the OSM Tasking Manager to coordinate mapathons. We self-host the tasking manager so that we can add functionality as necessary (e.g. custom iD editor links)
To stage data for import, we wrote 'osmizer' over the summer. The purpose of osmizer is to convert from a GIS input (GeoJSON) of pedestrian data that is in a semantic description (e.g. ‘type=sidewalk’) and output an OSM XML format with tags applied from the OpenSidewalks schema (‘highway=footway’, ‘footway=sidewalk’).
We've prototyped other data gathering apps. We've tried out POSM/OpenMapKit (not quite ready for our use case, but has massive potential), written our own passive data (gps, photos) gathering apps, and hybrid survey apps that combine GPS data with tagging (e.g. "I'm standing on a curb ramp").
A huge portion of AccessMap's infrastructure is how we go from Seattle open data to a pedestrian transportation network, because the municipal data is not even close to a routable state by default. The code is on Github, could be used by other cities. For Seattle, it does these things (in order):
Assigns sidewalk info (left/right/both/no and offset distance) to streets (because data.seattle.gov's sidewalks dataset has messy, incorrect lines and they need to be redrawn; sidewalk=left/right/both/no is also a common way for sidewalk data to be stored (including in OSM).
Redraws the sidewalk with parallel offsets using a standard algorithm.
Removes data errors caused by the parallel offset algorithm.
Algorithmically draws potential crossing locations between sidewalks at street intersections.
Assigns metadata to the pedestrian network made up by sidewalks and crossings (example: a crossing with curb ramps near both endpoints is marked as passable for those that require curb ramps).
Success stories
500-3000 unique users per month at accessmap.io, many more trips planned
Our mapathons have resulted in pedestrian data added to three Seattle neighborhoods so far, with many more in the pipeline
Opened up more opportunities for collaboration, including with the University of Washington to map its campus in preparation for the Special Olympics in 2018.
Further reading
Last updated