Good start. A UI improvement might include a change to the map, e.g. color coding, for the area around the address.
Also, please explain a bit more about how you compute walkability. My neighborhood only gets a 6%, probably because of a lack of sidewalks, but it is quiet with low traffic and therefore very walkable. So the score doesn’t seem to make sense for my situation.
This means that the score is considered relative to other areas in the US. A lack of sidewalks will negatively impact the score, for sure. Can I ask, how safe is it when there is traffic?
All of the above being said, if the score feels very wrong to you, it probably is! This will be a reflection of the fact that this is still at MVP stage
I built this with a friend to solve a problem: when people view properties to rent/buy, they often need to open Google maps in a separate browser window and type the address in again to find more information about the area.
Our solution:
Select an address and instantly get local information related to that address such as walking times and local amenities, without opening a separate browser window.
TL;DR
Walkability is an increasingly important aspect of our lives, so my friend and I built a tool to automatically measure it.
PROBLEM
* Walkability is important to people [0], and the planet [1]
* The existing solution, Walk Score, is flawed [2], seems to only operate as the crow-flies, and is missing key info, such as traffic safety [3]
SOLUTION
* My friend and I built an MVP to reflect recent academic literature [4] and provide a more detailed breakdown
* It’s an MVP to see what users think. The scores still need a lot of work.
EXAMPLES
1) 1000 Largo Center Dr, Largo, MD 20774, USA
Link: https://walkable.us/result/1000%20Largo%20Center%20Dr,%20Lar...
Image (in case site is down): https://i.imgur.com/cEik22U.png
2) Langley Park, MD, USA
Link: https://walkable.us/result/Langley%20Park,%20MD,%20USA
Image: https://i.imgur.com/6CiyFIa.png
IMPLEMENTATION
We reached a 100ms single CPU response time for any location within the US with the following:
* OSM data from GeoFabrik imported with imposm and pre-processed with osm4routing
* Data lives in a PostGIS DB with pgrouting and post_raster extensions
* Entirely route-network based walkability computation using OSM POIs, regions and edge data
* We achieved high speed in general by moving as much computation as possible to PostGIS - we were blown away by this toolchain!
* We use a PCA based model for the overall score, with range based breakdowns [5]
* Bing building footprints helped overcome OSM data paucity issues [6]
We hope you like it. Please let us know what you think!
REFERENCES
[0] https://www.bbc.co.uk/news/uk-england-london-56026940
[1] https://drawdown.org/solutions/walkable-cities/
[2] https://www.stateofplace.co/our-blog/2016/10/does-walk-score...
[3] https://www.sciencedirect.com/science/article/abs/pii/S13619...
[4] https://www.sciencedirect.com/science/article/pii/S209526351...
[5] https://www.semanticscholar.org/paper/A-Data-Driven-Framewor...
[6] https://www.microsoft.com/en-us/maps/building-footprints