A team of analysts have created Selfie City, a website to investigate the international socio-popular phenomenon of selfies. Lev Manovich, Moritz Stefaner, Mehrdad Yazdani, Dominikus Baur and Alise Tifen have examined tens of thousands of Instagram selfies from users in five different cities across the world and analysed their content using a mix of theoretic, artistic and quantitative methods.

Drawing from a pool of 120,000 randomly chosen selfies, the team used verification from two or more Amazon Mechanical Turk workers to filter this selection down to 1,000 photos from each city (New York, Sao Paulo, Berlin, Bangkok and Moscow). From the resulting set of pictures, the team ran an automatic face analysis, indicating algorithmic estimations of eye, nose and mouth positions and the degrees of different emotional expressions. Finally, a couple of team members manually checked through all the photos, omitting any that had been misidentified, and equalising the data sample size for each location. The final set contains 640 selfies per city, which can be explored on their site.

The Selfiexploratory section of the website allows you to examine these findings in an interactive data visualisation, which allows you to filter selfies by location, pose -- even degree of head tilt. As the Invisible Cities project gave a glimpse into residents online interactions, Selfie City allows us to better understand the variations within this global photographic past time.