This map shows the locations of 3.3 million geolocated tweets in the London area during the summer of 2012. That’s impressive in its own right, but the points are also colour-coded to show which language was used in the tweet (as determined by Google Translate).
The grey foundation of the map is formed from the majority of tweets, which were unsurprisingly in English. Many follow roads and train lines as people tweet while travelling. Other nationalities, in order of most to least prolific, are Spanish (white), French (red), Turkish (blue), Arabic (green), Portuguese (purple), German (orange), Italian (yellow), Malay (cyan) and Russian (violet).
It was put together by Ed Manley and James Cheshire at UCL’s Centre for Advanced Spatial Analysis (CASA) using tweets collected during the Olympic period this summer. In some cases, the clusters are pretty much what you’d expect — Arabic tweets centred around Edgware Road, for example. In other cases, the concentrations are less well understood, such as the large French enclave (red) in south-east London, or the French absence from South Kensington. More on the interpretation can be found on James’ blog, while further details on the underlying data are given on Ed’s site.
As the researchers note, the graphic is a highly skewed version of London, and shouldn’t be read as a reliable measure of the capital’s general demographics. It can only plot people with enough technological confidence to use Twitter, and the desire to use it. Still, it gives an elegant snapshot into London’s Twitter users in the summer of 2012. The work is based on the maps of Eric Fischer, who did something similar for the whole world.