“Hot or Not” for Cities


MIT from aboveMIT is full of cool labs.  SENSEable City Lab brought us the Copenhagen Wheel.  Now the Media Lab brings Place Pulse – an ongoing project to identify society’s perception of the safest, most affluent, and most unique places. The website is the front end of an impressive back end algorithm.  Geo-tagged photos of particular city scapes are presented to visitors of the site in pairs.  Users are asked to click which place looks safer, more unique, or more upper-class.  The algorithm uses the resulting binary data to classify places into these three categories.

That may not seem impressive, but check out the results.  Among all cities surveyed (Boston, New York, Vienna, Salzberg, and Linz), Boston and New York have the most unsafe-looking places, while the Austrian cities have the top 10 safe-looking places.  Just quickly looking at the photos categorized this way provides amazing insight into particular characteristics that draw people to, or send them running from, a public place. A sense of enclosure, vegetation, street-level retail, and other human beings are common characteristics among the safe-looking places.

As the developers note, these associations become even more powerful when mapped and correlated with other characteristics of their surrounding neighborhood or community.  These visualizations are not on the website, but will be published eventually along with a fully-featured, interactive website allowing anyone to participate in or create their own study.  If you’re in Austria, you should check out the exhibit in the meantime.

 

Comment