Ian Turton wrote:
The problems of cloud sourcing and geocoding information about Flu
outbreaks are hardly new - My student Andrew Murdoch gave a talk at
FOSS4G in Victoria (2007?) about it and we have presented else where
too - see
http://mapcontext.com/autocarto/web/documents/AutoCarto2008/Turton.pdf
for our latest paper.
One of the problems with the latest attempts is the use of twitter
where you get even less context to work with than the blog entries we
used.
I can completely picture a John Snow Cholera map that shows 15 reports
in houses around a pump due to poorl geocoding/book-keeping - where the
reality was a single incident, but falsely pointing to a clean pump.
This issue was a concern when we were working on TwitterVoteReport - it
would be too easy to have an issue appear blown out of proportion by the
crowd shouting louder (and therefore implying a more numerous problem)
than reality. This has been show more recently in the Mumbai attacks
regarding the retweeting that the gov't had asked people to stop
twittering, when it actually hadn't.
With VoteReport, we added a "Sweeper" interface. If you can crowd-source
the data gathering, then why can't you crowd-source the data curation?
Volunteers logged in and could "Approve", "Deny", or "Modify" (add
attribute data, wait times, correct the geocoded location) of any report.
The current work is on expanding out this platform in a concept coined
by Chris Blow (Meedan & Ushahidi volunteer) and Kaushul Jhalla called
"SwiftRiver" - http://swiftapp.org/ - There are increasingly large and
fast flows of information from trusted and public sources: articles,
twitters, photos, blog posts, SMS, et al. - they may be invalid, or 10
reports about a single incident, or be missing valuable attribute
information. The result are annotated, validated (or questionable)
stream of valuable information.
Erik Hersman has a good post about it here:
http://blog.ushahidi.com/index.php/2009/02/04/crisis-info-crowdsourcing-the-filter/
(and video here:
http://blog.ushahidi.com/index.php/2009/04/09/explaining-swift-river/)
Swift is in fact just repurposing the original VoteReport platform and
being integrated into the VoteReport.in (India) instance running
Ushahidi. The project is open-source
(http://github.com/ajturner/swiftriver/tree/master)
Anyways, don't discount the crowd - just situate one group against the
other :)
Andrew
_______________________________________________
Geowanking mailing list
[email protected]
http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
