Using more than 2 million social media images, we developed a model that quantifies the distinctiveness of 18 cities around the world and identifies their visual similarities. Select a city in the map to reveal its similar cities.
The model aims to predict from which city a given image comes. The percentage of images that the model correctly classifies from the city is its distinctiveness score. The most distinct images from each city are displayed below.
While each city has its own visual identity, they often share visual similarities with one another. We quantify the similarity between two given cities as the summed percentage of images that the model misclassified as the other city.
distinctiveness
score
The material on this website can be used freely in any publication provided that:
1. It is duly credited as a project by the MIT Senseable City Lab
2. A PDF copy of the publication is sent to senseable-press@bc.edu
For Contact:
senseable-contacts@mit.edu