roomthily

SalahEldeen and Nelson looked for tweets on six culturally significant events that occurred between June 2009 and March 2012. They then filtered the URLs these tweets pointed to and checked to see whether the content was still available on the web, either in its original form or in an archived form.

They found that the older the social media, the more likely its content was to be missing. In fact, they found an almost linear relationship between time and the percentage lost. 

via Technology Review

or the problems of social media as an historical archive

MappyHealth - data mining twitter for health term trends

prostheticknowledge:

Social Network Clock

Time-based visualization of social network activity based on bitly data, by retronator:

Today’s random idea to create a webpage: the Social Network Clock!

A few weeks back bitly posted some interesting data that hints when people use different social networks. Today’s stupid idea for not playing Minecraft all day involved creating a CSS3 time display webpage that overlays current time over the graphs published by bitly.

Tumblr is apparently the party network for evenings and weekends so cheers to you guys!

Link

LiveHood: New York City - social media and machine learning defining similar areas in a city (Carnegie Mellon University)

Sickweather - maps “illnesses” based on social media searches

topology, topology, topology. ick.

Social Flow’s Twitter visualization for #OccupyWallStreet in October

via The Why Axis

Google+ Ripples

via Visualising Data

Google+ Ripples - visualize how public posts are shared and commented

via Official Google Blog

Turf: Geography Club 

“foursquare meets monopoly” or where you earn coins for check-ins that can then be used to buy/upgrade property (real locations) or robots (not real robots)

via The Pop-Up City

Stories Unbound (Melbourne Writers Festival) - social media app to read and write geotagged stories