Another frequently voiced concern is the veracity of data collected on social media. It is a valid concern, but not the dead-end some formal aid actors assume it to be. The Social Media Emergency Management Camp report found that fear of incorrect data is a major cultural impediment to formal aid organizations’ acceptance of participatory media. But the teams building automated listening tools for use in crises operate under the assumption that some percentage of the data could be incorrect, or even downright malignant, and the overall collection of data will still provide a more complete situational awareness picture than responders have today. Or, as Meier puts it: “Perhaps we ought to keep in mind that even if only, say, 0.001% of the 20 million+ tweets generated during the first five days of Hurricane Sandy were actionable and only half of these were accurate, this would still mean over a thousand informative, real-time tweets, or about 15,000 words, or 25 pages of single-space, relevant, actionable and timely disaster information.”
Meier has written extensively about several emergent methods to verify crisis tweets and Josh Stearns has collected a wide range of tools and methods for verifying the information contained in social media. One robust example is SwiftRiver, a product in the Ushahidi family. The product seeks to help users manage the large volumes of data generated by social media with an interface that allows users to weight and filter the streams in real-time.
Meier has also detailed the large percentages of crisis tweets that contained useful, actionable information in a collection of past crises. The amount of actionable information the social media posts contain vary between crises and (significantly) across nations, but in every case, they offer novel, actionable information in some volume.
The Pew Research Center’s Project for Excellence in Journalism found News and Information to be the largest category of tweets sent during Hurricane Sandy, more than photos, jokes, prayers, or political commentary. Another forthcoming study analyzes Reddit threads discussing the storm, and finds that perspective-based imagery and other citizen reporting outweighed and outperformed links to professional news reports on the social news site.
The primary takeaway of social media’s contribution to situational awareness may be that the information it contains is still a major leap forward for situational awareness, even if the data is polluted with some inaccurate information. It’s important to note that most practitioners building social media listening tools and libraries do so in the hopes of augmenting existing needs assessment methodologies, not to outright replace formal methods with untested systems. These complementary systems can improve traditional needs assessments with real-time data and the improved geolocation of needs.