The Twitter Streaming API allows you to collect geolocated tweets and plot them on a map. With the new GeoEvent Processor extension for ArcGIS Server, it is now so easy to connect to this API and start harvesting this data. Esri Australia has already been making use of this for mapping out real-time coverage of Twitter trends during the recent election and mapping out the languages of Australia (plus mapping out profanity).
But a majority of the audience of this blog are likely GIS Professionals. Is there really any merit or use in mapping social media streams such as Tweets?
- Tweets have already been used in studies to try and better understand transport networks in the UK,
- When Natural Hazards occur, some people turn to Twitter for getting updates on whats happening. Esri have a template application that has been used for volcanoes, cyclones, flooding and bushfires.
- Can we predict riots by looking at Twitter?
- Can we determine the mood of a country by looking at Twitter?
- Not quite tweets, but mobile phone data is being used to map out bus routes in Africa
- Looking at Flickr, we can map out where we think locals and tourist photo hotspots are.
- Or maybe people just want a map to hang on their wall.
With all that said, its still only a fraction of the population that use Twitter, so is it really a fair ‘dataset’ to be basing decisions upon? For starters, only 5% of Australian internet users used Twitter in 2012. Less than 10% of Australian tweeters actually share their location. With the rise of mobile, these figures are increasing rapidy, and the typical age of a tweeter is also going up ( Twitter’s fastest growing age demographic is 55 to 64 year olds)
Either way, I like twitter maps. I put together a few useless fun maps with some tweets that I have been collecting with GeoEvent Processor. I best get back to some real work before my boss reads this post…
Click the maps to open in a new tab at higher resolution (some might take awhile to load up).
3 months of Tweets collected with GeoEvent Processor. Otherwise known as a population map…
10m buffer of tweets along Melbournes main busy roads during rush hour. Notice how a lot of people like to tweet at the traffic lights…
A clip of tweets within 50m of the main hipster strips within Fitzroy. Tag cloud shows most commonly used words in these tweets. (Naked for Satan is the bar to be at…)
Need to turn this into a web-map, its very interesting seeing the types of tweets people say as they are leaving/arriving an airport.
Esri Australia Technical Blog 