Harvesting Twitter's tweets through Web GIS for disaster relief

With technologies advancement, such as Web 2.0 and mobile devices, average citizens can contribute data based on their understanding to a problem, a phenomenon, or an event. Jeff Howe (2005) coined a new term “crowdsourcing”, which was later defined by Brabham (2008) as “online, distributed problem-solving and production model”. In crowdsourcing, the efforts from general public (voluntarily or without notices) or participants can help to solve problems or provide own their own about an event. The data from crowdsourcing actions has gained great attentions from academia researchers and industrial practitioners. One big category of crowdsourcing data are from social media websites, such as Twitter, Flickr. As user-generated-content, social media data provides real-information by volunteers from a local perspective which is not always realistic to have official agents on site during one event. Social media data has been used in national security, disaster management, social/crisis event, and other fields. In some climatic disaster events, such as Hurricane Sandy, social media data that were utilized by official agencies, such as FEMA and Red Cross demonstrated its power in locating victims, dispatching resources, and saving lives.

The advancements in technologies also enable new means of data collection. By utilizing APIs that are provided by social media websites, Web 2.0-based Web GIS portal can easily have users visually specify location for data harvesting, data mapping if applicable, and data storing. It can fuse collected, geo-tagged social media data with other geospatial data. The collected data can be saved in format as standalone files, such as CSV file, or can be saved in database for distributed access.

In the case of tweets from Twitter, this presentation majorly introduces Web GIS platform for real time tweets collection and live mapping of the geo-tagged tweets. By using Twitter's Search API, the web platform provides an interface for users to specify corresponding search parameters, including search keywords, search location, and other parameters, to retrieve subjects related and location related social media. Most importantly, it presents a way to save the real-time collected social media data into a distributed (geo)database, and disseminate the saved data lively.