We are back on again, and this time, what we are up to is indeed very exciting (not that others were not, but still). The SBTF is again in full activation mood: the Digital Humanitarian Network, which we have been one of the co-founders of, has been in fact activated by UNOCHA on 7 – Nov 2013 in response to Typhoon Yolanda.
The DHN Network has been mobilized to support Palau and the Philippines in support of the United Nations Office of Coordination for Humanitarian Affairs and the Philippines Red Cross to assist in providing information management support. In response to this activation the SBTF, Humanitarian Open Street Map, GIS Corp, ESRI Disaster response Program, Translators without Borders, Statistics without Borders, Info4Disaster and others have activated their respective volunteers.
The SBTF immediately called for the volunteers to start working in three separate teams: Clickers, Categorization and Geo-location.
With the support of the QCRI and their AIDR (still under development) and Twitris platform, the SBTF started immediately working on collected tweets in real-time that had been already filtered for both relevancy and uniqueness using automated algorithms. The ~55,000 tweets were processed using MicroMappers.
The cool thing about it, is that we are not alone this time: digital volunteers from the world can in fact now use this platform to tag tweets and images from the disaster.
Once the tweets have been categorized according to relevancy, the SBTF volunteers can process them again to:
1) Check that those tweets and images are actually relevant for responders
2) Add categories os that information can be sorted according to topic
3) Add location with GPS coordinates (longitude and latitude)
Once the data is cleaned, it feeds into this map and this one.
So, let’s also look at some numbers:
The automated system of QCRI harvested around ¼ million tweets
The Clicker system cut the number down to 17,932
After the SBTF volunteers worked on the data, the tweets that were left and ended up on the live map were 1,831 (around 0.3% of the Tweets harvested)
The SBTF volunteers that replied to the call were around 120
At the top of their activities, volunteers tagged around some 1,500 tweets in just 10 minutes
The SBTF team has an average of 8 people/hour working on the data
This is our first update, we hope to be able to continue updating you on this on a daily bases. If you want to see what we are doing step by step, you can see this Prezi.
For now, we would like to invite you to help us out by signing up here.
A huge Thank You to all the volunteers that are working hard to make this happen!!