LiveHoods: Using Social Media and Machine-Learning to Study Cities
April 19, 2012 11:17 AM Subscribe
LiveHoods: Using Social Media and Machine-Learning to Study Cities
Our research hypothesis is that the character of an urban area is defined not just by the the types of places found there, but also by the people that make it part of their daily life. To explore this idea, we use data from approximately 18 million check-ins collected from the location-based social network foursquare, and apply clustering algorithms to discover the different areas of the city.
We currently have LiveHood maps for NYC, SFBay area, and Pittsburgh. We're working on adding more cities. Would love to hear your ideas for how to improve the site, in terms of other kinds of analytics you'd like to see, top N lists (e.g. top 10 places most visited or top 5 places people travel the furthest to go to), places where the livehoods don't make sense to you, and so on.
Our research hypothesis is that the character of an urban area is defined not just by the the types of places found there, but also by the people that make it part of their daily life. To explore this idea, we use data from approximately 18 million check-ins collected from the location-based social network foursquare, and apply clustering algorithms to discover the different areas of the city.
We currently have LiveHood maps for NYC, SFBay area, and Pittsburgh. We're working on adding more cities. Would love to hear your ideas for how to improve the site, in terms of other kinds of analytics you'd like to see, top N lists (e.g. top 10 places most visited or top 5 places people travel the furthest to go to), places where the livehoods don't make sense to you, and so on.
Role: Faculty Advisor
This project was posted to MetaFilter by Potomac Avenue on April 20, 2012: The character of an urban area
« Older Ghosts With Shit Jobs... | Who Was David Algonquin? The W... Newer »
You've already added tons of interesting analytics (related livehoods, pulses, composition), but I have a couple suggestions since you're looking for them:
Perhaps you could include information about which check-in locations within each livehood are visited most frequently by individuals who most frequently check-in within other livehoods. That could provide a rough indication of what establishments are pulling in people from all over the city (and might be anchoring the activity of that livehood). I'm guessing it would probably match the top five places pretty frequently, but it might not always, so that could be interesting.
It would be cool to see a city overview. What livehoods city-wide are most active on each day and at each hour?
Also, how does composition and popularity change at different times city-wide and within different livehoods?
How exactly is the "Top five unique things to do here" calculated? You might consider adding a blurb in about that.
Finally, I have to throw my vote in for Long Beach, CA next, because I live here. And, as that's probably a long shot, my second vote would be for LA.
posted by Defenestrator at 1:40 PM on April 22, 2012