“How to establish trust in citizen reported human crisis reports” is my PhD question. So far. Ushahidi reports on ‘crisis’ (need to define that), in the vague field that is crisis mapping. I’ve had a few ideas which I’ve spoken at WhereCampEU, namely on the classification of crisis and trust in the reports of events on the ground.
In this I think that there are four categories of crisis, across the axes of severity (acute to severe) and time (a few seconds to ongoing indefinitely) – On presenting this at WhereCampEU a comment was made I was waiting for the good bit of news, then I realised it’s all bad! Most of the attention from the community is, seemingly, in mapping at midlevel to severe crisis over a medium time-scale. Examples of these sorts of crisis would include the Kenya election of 2007 and the Arab spring uprisings of 2011 in Tunisia and Egypt; something happened in a few weeks the crisis was ‘resolved’ – I’m not suggesting that the problems are solved, but they’re not in a state of active warfare.
Topics like earthquakes and floods would fall into roughly the same category as the uprisings with critical immediate effects, whereas bombings and its ilk would be severe but over quickly, also the spatial distribution of reports probably wouldn’t be as great as some disasters like civil wars.
I believe it also important to stress the whole process isn’t about the map creation process, a map is a representation of the world, not the physical one. I’m increasingly feeling that though the outside community response tracing roads and POIs from map tiles was invaluable in Haiti, I feel reports (SMS/Twitter/Web) generated by the crowd within the crisis zone is of better value. From this a context map of what is going on in certain areas could be raised, though I’m still considering how this can be visualised, and fit into a useful chain of use.
From this rambling, over the next six months I’m aiming to do some statistical experiments on crisis data, looking at which mediums people use to report, the percentages of which are verified based on input method (SMS/Twitter/Web/etc). I would also like to see a definitive case study of the ‘supply chain’ of information, from the user reporting on the ground to an organisation that uses that report. I believe understanding how useful the reports are to the aid organisations would be a good start on understanding this complex system.
In all, please comment on the ideas presented, or if you’ve seen information around the web or in journals, I would be very interested for similar people to be in touch.