Current Issue
This Month's Print Issue

Follow Fast Company

We’ll come to you.

1 minute read


Visualizing London's Skyline With A 3-D Map Of Tweets

Some people say a city can't be built in a day. But it took these researchers only 16 hours worth of tweets to grow the London skyline from the ground up based on messages geotagged to specific buildings.

Visualizing London's Skyline With A 3-D Map Of Tweets

We've seen how 3-D mapping can be used to visualize a city's history, or even more abstract concepts like relative levels of income inequality. But with all the geocoded data floating around the social media sphere, there are a growing number of quirky questions just waiting to be asked.

Take the two doctoral researchers at London's Centre for Advanced Spatial Analysis, who wondered if they could map the city's skyline by the volume of building-specific tweets.

More than nearly 16 hours and 3,500 London-based tweets later, Stephan Hugel and Flora Roumpani grew the city from the ground up using CityEngine, a tool for generating 3-D urban landscapes, the Twitter API, and a Python program they built that harvested tweets from the area. Hugel and Roumpani demonstrated the result in the video above, in which you can see city landmarks like the Marble Arch and Waterloo assume the size of their social media commentary.

Quite rightly, the pair were careful to note that this preliminary mapping experiment was just that—a proof of concept, rather than a display of solid research. "Ideally, rather than showing 24 hours of data, you could show a week, or a month, then track the peaks and troughs," Hugel told Co.Exist. "We don't want to stop with Twitter data, but for our uses Twitter data was the lowest hanging fruit," Hugel added.

Still, even in its infancy, Hugel and Roumpani's Tweet City could go on to demonstrate a different kind of city inequality. If they were to apply the model to New York City, for example, it would be interesting to see if the areas with little to no geocoded data also correspond to neighborhoods with slow Internet connection speeds or fewer Internet-connected devices. "In most cases, you would be showing off the fact that the data you're generating is by people who can afford to do so," Hugel noted.

But while geocoded tweets provided easy access to mappable data, Hugel and Roumpani are looking to expand the type of information they use. Currently, they're looking to grow buildings by air-quality readings, though tracking down enough air-quality sensors to form accurate, pinpoint pollution estimates is a challenge. In the meantime, Tweet City is available for anyone to download and fiddle with on Github.