Last Thursday, I attended a webinar that was called Monitoring Humanitarian Crises in the Digital Age: Crisis Mapping, Crowdsourcing, and Satellite Imagery. The webinar was hosted by Harvard University’s Program on Humanitarian Policy and Conflict Resolution. Dr. Patrick Meier of the Qatar Foundation’s Computing Research Institute (QFCRI) led much of the session. He spoke about crisis mapping and big data in the digital age. This article summarizes some of Dr. Meier’s discussion during that webinar.
Last Thursday, I attended a webinar that was called Monitoring Humanitarian Crises in the Digital Age: Crisis Mapping, Crowdsourcing, and Satellite Imagery. The webinar was hosted by Harvard University’s Program on Humanitarian Policy and Conflict Resolution. Dr. Patrick Meier of the Qatar Foundation’s Computing Research Institute (QFCRI) led much of the session. He spoke about crisis mapping and big data in the digital age. This article summarizes some of Dr. Meier’s discussion during that webinar.
Crisis Mapping Defined
Crisis mapping is the use of real-time crowdsourced crisis event data, satellite images, data visualization, data modeling, and web-based applications to develop early warning and response systems for use in crisis events world-wide. Crisis mappers perform big data analytics and data mapping in order to glean insights about what and where crisis events are occurring on a real-time basis.
Big Data and Crowdsourcing in Humanitarian Crisis Mapping
The Use of Artificial Intelligence in Analyzing Big Data
During Hurricane Sandy, 20 million tweets were posted about the disaster in a 5 day period. During the 2011 9.0 earthquake in Japan, approximately 18 million tweets per hour were posted about the disaster. The volume, velocity, and variety of this type of data stream are staggering, but the data must be analyzed in order to gleam insights about what and where crises are occurring.
The volume of these data sets is one tremendous problem. Crowdsourcing and microtasking aren’t capable of handling this sort of data load. To derive insights from this volume of data, artificial intelligence tools such as natural language processing and machine learning must be implemented.
Another problem with using crowdsourced real-time data streams from Twitter is found in determining the credibility and actionability of the data generated. As part of Dr. Meier’s work at QFCRI, he and his team are working on systems to automate the analysis of socially-derived big data to determine its credibility and actionability.
Some Major Players in Big Data Crisis Mapping
The Digital Humanitarian Network (Tw- #DHNetwork) is a network of networks that hosts volunteer and technical communities that are available to help with real-time social data monitoring, big data and GIS analysis, geo-referencing of event data, and crisis map production. Additionally, there is Crisis Mappers (Tw- @crisismappers) – a humanitarian technology network that focuses on crisis mapping for humanitarian support efforts.
Members of the Digital Humanitarian Network
“So”, you may ask, “who are the users of crisis maps?” To name just a few, the UN Office for the Coordination of Humanitarian Affairs (UN-OCHA), the American Red Cross, USAID, and even public for-profit entities such as the Washington Post are all utilizing crisis mapping during disaster events.
Crisis Mapping in Typhoon Pablo
Days before Typhoon Pablo made landfall in the Philippines, the Philippine Government began to inform its citizens about what Twitter hash tags to follow and tweet during the emergency event. After the typhoon hit land, the UN Office for the Coordination of Humanitarian Affairs (UN-OCHA) activated the Digital Humanitarian Network and tasked the Network’s stand-by volunteer task force with monitoring and analyzing all tweets generated within the first 2 days of the event. This analysis had to be completed and results submitted to the UN within 12 hours.
The taskforce utilized crowdsourcing and the Pybossa microtasking platform in order to isolate and analyze over 20,000 tweets that provided information and video footage about damage that was done during the disaster. The results of this analysis were then taken by UN-OCHA and used to generate the crisis map that is shown below. This is the first-ever UN crisis map that was generated solely from social media data. It was used as part of the coordinated damage assessment support that was led by UN-OCHA.
Photo Credit: Monitoring Humanitarian Crises in the Digital Age: Crisis Mapping, Crowdsourcing, and Satellite Imagery Webinar by HPCR at Harvard
Are you a big data analyst that is involved in crisis mapping? Do you know of any? If so, what sorts of humanitarian support applications have you or your friend developed with your skills in big data analytics?