Chris Orwa from iHub Research in Nairobi discusses the Deployment of Machine Learning During the Kenyan Election. Social networks are awash with information. Relief agencies such the Kenyan chapter of the Red Cross are already leveraging information from Twitter to track and respond to emergencies. The innumerable amount of information generated within a crisis requires faster processing to extract actionable information. At iHub Research, we studied the flow of information on social media during the March 2013 Kenyan general election and developed a framework looking at the '3Vs of crowdsourcing,' a functional approach to validating, verifying and checking viability of crowdsourced information.
As part of the research, machine learning techniques were deployed to sift through 2.6 million tweets and remove non-pertinent data, which narrowed down to 12,000 useful tweets. Further clustering of the data obtained 98 unique incidences. The techniques deployed are fundamentally useful to digital humanitarian volunteers for processing crisis data. Chris Orwa, a data scientist at iHub Research, will give a talk on the use of machine learning techniques for online crisis response. @blackorwa