Wednesday, July 09, 2014
Routing for Beauty
Last month Google Maps Mania reported on MIT's StreetScore. StreetScore rates the perceived safety of locations in New York, Boston, Chicago and Detroit based on crowd-sourced reactions to Street View images. The ratings are based on MIT's citizen science project Place Pulse.
StreetScore gives a score to each Street View location to rate how safe people perceive that location based on the overall responses to its Street View image. Over the last month I've been thinking a little about how the StreetScore data could be used in a routing application to suggest routes based on perceived safety. It turns out that like all my great ideas someone has already beaten me to it.
UrbanGems has been running a similar citizen science project as MIT's Place Pulse. Urbangems shows users two Street View images side-by-side and then asks them to vote on which one is more beautiful (or quiet or happy). UrbanGems is then able to rank how beautiful (or calm or happy) each location is perceived based on the combined votes of all the participants in the project.
UrbanGems has now released a research paper, 'The Shortest Path to Happiness: Recommending Beautiful, Quiet, and Happy Routes in the City'. In the paper the researchers propose a mapping service that automatically suggests routes that are perceived as more beautiful. The routing service would use the ranking scores derived from the UrbanGems project to suggest routes that may add a few extra walking minutes to a journey if it is perceived to be more beautiful (or quiet or happy).
If you want to create your own routing application based on perceived safety you could use the data from MIT's StreetScore project. MIT has released the data of the StreetScores for Boston and New York and it is available for download in two comma separated CSV files.
It is a fairly simple task to create a ranking for each block section of a street by combining all the StreetScores between two intersections. Once you have rankings for each block section of all the streets in a city you could create an A* search algorithm to search for routes that are perceived as more safe.