Mapped: Twitter Languages in London


[zoomit id=”IIY6″ width=”auto” height=”400px”]

**Update: You can see a new fully-interactive version here**

Last year Eric Fischer produced a great map (see end of post) visualising the language communities of Twitter. The map, perhaps unsurprisingly, closely matches the geographic extents of the world’s major linguistic groups. On seeing these broad patterns I wondered how well they applied London- a great international city. The graphic above shows (and here for non-zoom version) the spatial distribution of about 3.3 million geo-located tweets (based on GPS) coloured by the language detected using Google’s translation tools. Ed Manley collected the data and he goes into more detail about the data here. They cover the summer period so we can clearly see the many languages of the Olympic Park (a hotspot for tweeting). English tweets (grey) dominate (unsurprisingly) and they provide crisp outlines to roads and train lines as people tweet on the move. Towards the north, more Turkish tweets (blue) appear, Arabic tweets (green) are most common around Edgware Road and there are pockets of Russian tweets (pink) in parts of central London. The geography of the French tweets (red) is perhaps most surprising as they appear to exist in high density pockets around the centre and don’t stand out in South Kensington (an area with the Institut Francais, a French High School and the French Embassy). It may be that as a proportion of tweeters in this area they are small so they don’t stand out, or it could be that there are prolific tweeters (or bots) in the highly concentrated areas. I really like the paint-speckled effect that the multilingual tweets of London have produced and it offers a further confirmation of the international nature of London’s population.

Even though the map contains over 3 million tweets it is still a fairly selective sample of Londoners- they only include people who have a good location (through GPS) and those who are connected to the internet. I expect the latter requirement will exclude many short term visitors to London, and may explain why there aren’t so many hotspots around London’s landmarks (as is the case with Flickr where people can upload georeferenced images when they get home). There are also a couple of horizontal lines that have been caused by different levels of precision in the tweet locations. In spite of this, I think the information in these maps is useful as a basis for comparison to other cities and it helps to reveal some of the finer patterns within the broad regions mapped by Fischer.


  1. Carl Thomas

    I saw your map in the Metro and was interested in the French language clusters. In particular the one towards bottom right stood out and as far as I can figure out is Lewisham. It certainly isn’t what I would expect but perhaps we’re seeing a large number of African French language tweeters?

    1. Tom

      I think this may be Blackheath and surrounding area. French Saturday school and many French families who have moved in the last 4-5 years.

  2. Andy

    Really enjoyed reading this article – first saw it in the Metro this morning. Can you write a step by step guide on how anyone can do this sort of production? Would you be up for meeting in person re this? I’ll buy you a drink! 🙂

    Many thanks,

    1. Ed Manley

      Hi Andy,

      The tweets were downloaded using the Twitter API, for the specified bounds of London. There is a fair bit of documentation online describing how to set this up.

      We then used the Chromium Compact Language Detector to identify the language within each tweet. This is a Python library and freely available here – This code was simply included in a Python script that looped through the tweets and output the language. A reliability indicator is also provided, in our case we only used ‘reliably detected’ results.

      There is a bit more information on the process and results on my blog –

      Let me know if you have any further questions.


      1. WillB

        Hi Ed,

        This is a great map. What graphics/visualisation tool did you use to map the data programmatically? Im trying to find a good programmable tool that I build complex data visualations in.



  3. Tristán White

    Really enjoyed this, though would have loved to be able to get better quality on the zoom. Also, is there anywhere a list of all the languages found?

  4. mad tony

    very ineresting…more importantly thanks for shpwing the full data and the high proportion of unknowns. Google translate is good but not foolproof

    Love the maps in particukar, it is a shame it could not have another map superimposed on it

  5. Julian Boyle

    This is a beautiful map. I am just wondering what language the Chinese community are Tweeting in – they (as the second largest of all language groups) appear to be absent from the map?

  6. Veriest

    Julian Boyle (and to the authors): there is a Chinese equivalent of Twitter called Weibo, which is probably why there is no Chinese showing up at all. I’d imagine that most all users who tweet in Chinese would do it on Weibo instead of Twitter.

    These infographics are beautiful!

  7. This map is extremely highly biased. This is for exampel extremely suprising to see that Tibetan is more represented than Hindi, given the very visible Indian community in London (and I doubt they will use English instead of Hindi more than what Tibetan use, notably in Tweeter where there’s a lot of difficulties to type that Tibetan language !!!)
    For this reason, it will be much more relaiable if using posts on FlickR, where people are gelocalising their photos and uploading them, before commenting them on the web in a more reliable way in their own language, and where other people are commenting the photos.

    I don’t think that Tweeter is a good indicator of languages used! It will be highly biased in favor of English or languages that thare natively supported on all mobiles phones (because they all support the minimum Latin-only character set, and partly supported in Greek and Cyrillic scripts by reinterpreting Latin letters as Greek or Cyrillic, because these two scripts are NOT encoded separately on GSM, or simply not supported by SO many mobiles phones and even MANY “smart” phones, including when using the Tweeter application instead of SMS messaging to post tweets)

    And yes, the **heavy** costs of mobile Internet when roaming abroad forbids many people of posting tweets when they are abroad, unless they actually live there for a long enough time to get a residence and a local internet access with affordable (or flat) mobile data plans. One day Europe will require ISPs to provide pan-European data access plans at no additional cost.

    Or may be we’ll finally see free Wifi access
    everywhere, at least in very urbanized areas like London (but if you’re limited to only urbanized areas, then you have also another bias: the spots you see on the map are also those where there are FREE Wifi hotspots (notably restaurants and cafés) for visitors.

    So you should study this map differently across zones of London : areas with restaurants, cafés, hotels that offer free Wifi access will be MUCH less biased. Transportation stations ARE biases because the Wifi hotspots are NOT free even if they are MUCH less expensive than mobile Internet proposed by roaming mobile phone operators (most hotspots in transportation areas are operated commercially by BT, Vodafone, Orange, … they are definitely NOT free, even for reataurants and cafés located there that can’t open their FREE Wifi Hotspot).

    It would be interesting to look at tweets posted from **McDonald restaurants** (its local hotspots are free everywhere, and its presence worldwide gives good comparison points for creating statistics about languages spoken in these areas by BOTH visitors and residents).

  8. Adam Crymble

    This is really cool. Although I’m a bit worried mapping where the foreigners are could be used for ill-means by xenophobes.

  9. Neil Appleton

    Hi Ed, James – this map is fantastic. I’d love to stick this on my wall at the highest resolution possible. And would be very happy to pay for it :-). Can you recommend a next step! Best wishes (and congratulations on a beautful visualisation), Neil

Comments are closed.