Caveat, this has heroku-rotted and probably doesn’t work anymore. One day I’ll get this up and running elsewhere.
Los Angeles Linguistic Geography is an experiment in visualizing real-time linguistic data. The focus of this project is to explore the possiblities of real-time local visualization of information and communication within a city.
Data on tweets in languages other than English are visualized based on their location. As the tweets occur in real-time, they are displayed and pop up on the map. Each color on the map represents a different language, according to Twitter’s data.
Map data is sourced from Esri, using the Esri Leaflet plugin.
Linguistic data is sourced from Twitter, and is embedded within Tweets that are accessed via the Twitter API. The server subscribes to all tweets within the bounding box of Los Angeles, and filters to only store/display Tweets that either are tagged as in a language other than English, or whose users are configured as having a native language other than English.
It turns out that Twitter is quite bad at detecting language, to great humerous effect during my presentation. The next steps would be to filter out obvious false positives, then begin to do analytics on the data, to give more up-to-date information on language use in the community on this particular social network.
Source can be found on github.