Design Principles for Information Categorization Quality in Crowdsourced Crisis Mapping Platforms
研究了2010年海地地震中Ushahidi平台的案例，提出设计原则帮助众包平台将受害者提交的混乱求助信息转化为结构化的危机报告，并通过实验验证了模板对提升分类质量的有效性。
Crisis mapping platforms have transformed disaster management and digital humanitarian efforts by allowing victims to quickly submit “Requests for Help” (RFH) messages directly from disaster locations via mobile devices. On these platforms, online volunteers are often engaged in processing and categorizing messy and incomplete RFH messages into structured and useful crisis reports that can aid first responders in their recovery efforts. This research note examines the case of the Ushahidi platform deployed during the 2010 Haiti Earthquake to propose design principles for a crisis mapping platform that facilitates the conversion and categorization of victim-submitted RFH messages into actionable crisis reports for on-site first responders. To validate the proposed design principles, we instantiated them with the help of a template, and conducted a series of experiments to confirm the effectiveness of the template in improving the categorization quality of crisis reports. We expect that the design principles will be particularly useful for developing digital platforms aimed at humanitarian crisis response that requires a large-scale participation of online crowd volunteers.