Data science meets social science: using social media to understand disaster response in Australia

Australia is hit hard by all sorts of natural disasters, both in regional (e.g. bushfires) and urban (e.g. floods) areas. The economic and social damage from natural disasters is increasing as they become more frequent and as more people live in areas of high risk. The use of social media in crisis situations has emerged as a useful modern tool to improve disaster preparedness and management.
Twitter, a widely used social media platform, can disseminate information in an emergency faster than traditional channels and is widely employed to warn and inform people. Affected people also communicate among their networks about their needs during a crisis. This communication through Twitter tweets can be analysed using machine learning based techniques designed to analyse big data. Such an analysis can improve the understanding of what is happening before, during and in the aftermath of natural disasters, knowledge which can be used to reduce the human and economic impacts of future disasters.
While social media messages sent in natural disaster situations has been the subject of research outside Australia for over 15 years, their potential has not been realised for Australia. The aims of the proposed project are to:
• explore what kind of useful information social media data can be collected from communications before, during and after disasters can provide
• explore for which disasters this kind of data is the most relevant
• explore how the data can be analysed and compiled in a manner useful for policy-makers and stakeholders working in disaster risk reduction