A new tool for collaborative social media analysis in disaster response
During large-scale complex crises such as the Haiti earthquake, the Indian Ocean tsunami and the Arab Spring, social media has emerged as a source of timely and detailed reports regarding important events. However, individual disaster responders, government officials or citizens who wish to access this vast knowledge base are met with a torrent of information that quickly results in information overload. Without a way to organize and navigate the reports, important details are easily overlooked and it is challenging to use the data to get an overview of the situation as a whole.
We (researchers at Madeira University, IBM Research and University of Oulu) believe that volunteers around the world would be willing to assist hard-pressed decision makers with information management, if the tools were available. With this vision in mind, we have developed CrisisTracker.
Automated story detection
CrisisTracker offers an alternative way to browse social media activity around large-scale events, in particular disasters. The system automatically tracks a set of keywords on Twitter and clusters tweets based on their word similarity. Information propagates on Twitter mostly through duplication, either directly through retweets or by multiple people independently talking about the same event, and information about a single event is often split on thousands of messages. By clustering messages, each piece of information (a "story") is decoupled from its sharing pattern. Once decoupled, information about who and how many people shared a piece of information can be used to estimate how important the information is, and to whom, without requiring a computer to first understand what the information is about.
Crowdsourced information management
The web interface supports exploration of stories as they unfold in real-time; by time, location, topic and named entities. As current state-of-the-art text processing algorithms struggle with reliably extracting such meta-data from the text in tweets, CrisisTracker instead explores the use of crowdsourcing techniques. Any user of the system can directly contribute tags that make it easier for other users to retrieve information and explore stories by similarity. In addition, users of the system can influence how tweets are grouped into stories.
What is the difference between CrisisTracker and Ushahidi?
The biggest difference between the platform and Ushahidi is that Ushahidi focuses on curation of user-submitted reports, while CrisisTracker mines Twitter for reports, clusters them, and supports curation of report clusters. Both systems require humans to annotate pieces of information with meta-data such as location and report category.
This means that an Ushahidi deployment generally handles much fewer but higher quality reports, as each report is written specifically for the platform. CrisisTracker on the other hand can collect information even if sources are not aware of the deployment of the system as it listens in to a publicly accessible communication channel (Twitter) that people already use to share their knowledge in crisis.
The purpose of CrisisTracker is to create overview of what is being said in social media (including rumors and bias towards specific topics) more than to create a perfect model of the real world. It helps analysts find where reports originated and to get in contact with sources (Twitter accounts). It also helps in finding similar but potentially conflicting versions of a report. CrisisTracker's contribution is greatest during complex large-scale events when it bring order to overwhelming social media feeds consisting of hundreds of thousands or millions of tweets over many days. Other tools are more suitable for monitoring small-scale localized events that produce only a few hundred tweets in total.
We are extremely interested in hearing your feedback! Just use the tab on the right, or visit our discussion forum.
Jakob Rogstadius - Research, design, development (M-ITI)
Claudio Teixeira - Web development (M-ITI)
Maja Vukovic - Advisor (IBM Research)
Vassilis Kostakos - Advisor (University of Oulu)
Evangelos Karapanos - Advisor (M-ITI)
CrisisTracker is free and open source, so that you can deploy your own instance or integrate it with your own analysis software.
CrisisTracker in the news
Crowdsourcing Platforms Monitor Disease, Dissent, and Disasters
Communications of the ACM, 8 March, 2013
As Prop, Cudgel or Sensor, Digital Maps Have a Future in Global Activism
TechPresident, 16 January 2013
Crisis Tracker: An Open Source Map that Curates Crowdsourced Information
TechPresident, 1 November 2012
Social media in emergencies: is there a responsibility to verify?
Social Media 4 Good, 20 September 2012
New Social Media Monitoring Tool: CrisisTracker
idisaster 2.0, 17 September 2012
Crisistracker, nueva herramienta para organizar en internet la información sobre catástrofes
el Periódico, 21 August 2012
CrisisTracker: Collaborative Social Media Analysis For Disaster Response
iRevolution, 30 July 2012
Earthquake? Terrorist bomb? Call in the AI
New Scientist, 23 May 2011