As you may have noticed, I’ve been slacking off in my blogging duties recently (not withstanding yesterday and the amazing guest posts you saw over the last couple weeks). While I offer my sincerest apologies, I promise there was a perfectly good excuse: I was busy getting married and then relaxing on the beach with my new husband.
Okay, maybe that’s not the best excuse but perhaps it’ll make you feel better to know that I’m finally back in the game. My head is once again screwed on straight after months of worrying about things like the color of plastic chair our guests’ bottoms should be stationed in and whether or not to endeavor into the folding of 90 paper cranes as seating cards (I opted in, despite formerly-referenced new husband’s protestations).
Returning to work after even a short hiatus is always daunting, if only for the fury of emails you find yourself sifting through. But this time, after I got through all the Google alerts, I felt a little like a kid in a candy shop: shelves full of wonderful scientific treats, mine for the blogging. One of these came from network scientist Alan Mislove, an assistant professor in the College of Computer and Information Sciences.
A few years ago I stumbled upon a map of the US by artist Ben Fry that depicted nothing but streets. Geographical features like mountains, coastlines, rivers, and valleys emerged as a result of increasing or decreasing thoroughfare density. Likewise cities became visible as dark concentrated nodes, clots of tangled streets.
Mislove has something similar but, astonishingly, way cooler. It’s an interactive map of the entire planet with nothing but Tweets drawing its lines. That’s right, Tweets. From Twitter. The visualization is created from the 275 million geo-tagged tweets from a data set of more than 10 billion that Mislove’s team collected between January 2011 and April 2013.
Just as with Fry’s All Streets map, Mislove’s “all Twitter” map reveals a system of streets and roads that tell a geographical story.…What’s amazing here is that those streets are emerging from the Tweets of the people populating them. “The resolution of the data, the fact that you could see roads, etc., blew me away,” said Mislove, who did all of this as a side project just to visualize where on the planet his data points were coming from.
The map also shows the ferry lines running through the English Channel and across the Irish Sea, for example. Or between Playa del Carmen (where I just spent ten days drinking pina coladas) and Cozumel. But the big thing this map shows us is where people are Tweeting most (or at least where people are geotagging their tweets the most). Unsurprisingly, the US, Europe, and Japan take the cake, but so too do the coastlines of nearly all the continents.
For future work, Mislove’s team is looking at how Twitter usage varies across different regions. In the map below, the different colors represent the predominant language spoken at each location (orange=English, blue=Spanish, green=French, red=German, purple=Portuguese, pink=Arabic, black=other).
This reminded me of some cool work from another network scientist Alessandro Vespignani, Sternberg Family Distinguished University Professor of computer and information science, health sciences, and physics. In a project called the Twitter of Babel, Vespignani and his team used geo-localized Tweets and automatic language detection to map the most commonly used languages among Tweeters around the globe. As you can see below, the picture, though not as colorful as Mislove’s, tells a similar story. You can scroll over Vespignani’s Twitter of Babel map to see the breakdown of specific languages spoken in each region. Does it surprise you as much as it does me that there are more Tagalog-speaking Tweeters in Pennsylvania than there are Spanish-speaking ones?