A new crowdsourcing model to assess disaster using microblog data in typhoon Haiyan

Risk prediction and damage assessment play critical roles in disaster response to reduce losses. Social media can serve as crowdsourcing platforms for disaster information dissemination and data mining. Using typhoon Haiyan as an example, a close relationship between social media and disaster damage estimation is demonstrated, which provides a new perspective for disaster preparedness and response. Based on disaster-related social media data, a new index model is developed for situation awareness and damage assessment before, during, and after disasters. The difference between the new index model and traditional ones is that the new index is extracted from microblogs using semantic analysis method. The score of each index is determined by the emergency management experts. The weight is calculated based on TF-IDF method, a classical term frequency weight method. Based on the new index model, quantitative assessment is added to qualitative analysis. The assessment result is consistent with actual situation, which underlines the feasibility of implementation of the new model.