What it does
It takes what people all over the world are posting on Twitter with the hashtag #nowplaying (last 500 tweets) and tries to extract the artists played using some heuristics.
Using artist similarities measures (Universal APIs) it computes a target artist that would be a good compromise for the music tastes of the people, in that moment, considering what they posted on the last 500 tweets.
It plays a TOP song by this target artist (Spotify APIs) and at the end of the song the process re-starts, analyzing again the last 500 tweets, computing the target artist and playing a song.
At the same time, every two seconds a graph representing the "world beat" is refreshed: the app shows an equalizer representing the mix of genres/styles played in that moment and the inferred people aggregated mood. The aggregated mood is computed assuming that people play a more lively song when they are in a good mood and it's computed using the valence  audio feature provided by the Spotify APIs for the song played by the radio.
 A measure from 0.0 to 1.0 describing the musical positiveness conveyed by a track. Tracks with high valence sound more positive (e.g. happy, cheerful, euphoric), while tracks with low valence sound more negative (e.g. sad, depressed, angry). See link