Some editing stats from the Typhoon Haiyan response
On Friday I presented a talk at the Open Data Institute on "Typhoon Crisis Mapping With OpenStreetMap". You can see the slides and notes at that link, and even an audio recording. The talk was an intro to OpenStreetMap and H.O.T.
Maybe you've heard all that before, but I also tried to look back on our recent efforts responding to Typhoon Haiyan in the Philippines, and I thought I would share these graphs of editing traffic here:
The graph shows editing traffic over time. The time of the Typhoon is on the far left. So you can see we had a flurry of activity in the weeks following the disaster, which has tailed off. This corresponds very closely to the level of interest in the disaster in general (In my slides I compared it to google trends for the word “Typhoon”, which shows the amount of press coverage and buzz about the disaster) Unfortunately we only get lots of volunteer enthusiasm when these things are in the news, although I think this may also show we have a slightly longer attention span than the rest of the internet!
The spike isn’t a bad thing though. In fact in some ways we need to work on making the spike taller, but shifting it to the left as much as possible. The best and most effective response mapping happens as soon as possible after the disaster. Imagine an aid organisation using the data, by taking a snapshot from OpenStreetMap 48 hours after the disaster. The bulk of edits will not be picked up by them.
This shows new users in green. These are people who appear to have registered to edit OpenStreetMap during the crisis response (probably purely in order to take part in the crisis response).
This maybe shows that our “old-timers”, the more experienced OpenStreetMappers, were a bit quicker off the mark, responding soon after the disaster, while the new users came along when we ran events and showed them how to do it.
I was able to throw those graphs together fairly quickly thanks to Pascal Neis, who gave me his data he had captured (filtering for edits in the area) while making his changeset visualisation. OpenStreetMap has very rich meta-data on who edited what and when, so clearly there's more analysis data-mining of this kind we can do.
Whether you were a "new user" or an "old timer" on this graph, well done to everyone who helped map the Philippines in response to typhoon Haiyan. After my talk, during the questions, H.O.T. received warm endorsements from British Red Cross, MSF UK, and MapAction who were all in attendance.