CrisisNET speedily aggregates social data in disaster situations (Wired UK)


Ushahidi/CrisiNET


Advertisement

Advertisement

Not-for-profit software company Ushahidi
has launched CrisNET
, an open-source platform that it claims
will dramatically reduce the amount of time that it takes
journalists, analysts and humanitarian organisations to get their
hands on well-structured, crowdsourced data in the midst of
conflict and disaster.

Ushahidi describes CrisisNET as a “the firehose of
global crisis data”
. It has been designed to suck in data from
a number of sources, then consolidate the data into a useful format
to help people effectively monitor a crisis situation. The
organisation has a track record of helping to gather data from
social media during natural disasters and political revolutions –
it was formed in the aftermath of the disputed 2007 Kenyan
presidential election — but the sheer quantity that it now pulls
in is expensive to clean and format at speed.

Working together with crisis blogger Eliot
Higgins
, who has been closely documenting the unrest in Syria
using social media reports, Ushahidi tested CrisisNET by pulling in thousands of Facebook
and YouTube pages, as well as information published by media and
non-profit organisations to create a single stream of continuously
updated, real-time data. These sources of information would have
otherwise required hundreds of hours of manual process to determine
what was happening in Syria.

Ushahidi automated the monitoring and analytical activities that
Higgins usually would complete manually, leveraging many of the
same fundamental techniques that search experts like Google used to
find patterns in search history or identify hashtag popularity. In
order to train the system to recognise things like relevance,
topic, names and places, they had to teach it recognition skills
that worked on a par with the trained expert human eye.

After first translating all content into English, then scanning
for keywords and phrases, CrisisNET has to employ complex natural
language processing in order to identify location information
within an update, image or video that it can then add to the
metadata. From here, geotagged data can used to create a map.


Ushahidi/CrisiNET


Machine-aggregated social media reports created by CrisisNET
have been found to clearly correlate with the situation as
documented by journalists and organisations on the ground –
Ushahidi demonstrates this by comparing a CrisisNET map to one
created by the BBC according to reports from the scene — proving
its value as a tool for those reporting on crisis situations.
“Social media and data-driven analysis are the journalist’s new
secret weapon in monitoring conflict,” writes the CrisisNET team in
a blog post.

Advertisement

Advertisement

CrisisNet is now in public beta and is being provided free to
everyone that needs access to crisis data quickly.

10 June 2014 | 12:29 pm – Source: wired.co.uk
———————————————————————————————————————

Leave a Reply

Your email address will not be published.