Week 10: 7/24-7/30 Special Topics: Crisis/Disaster Informatics
Crisis informatics
As this week’s video materials and readings make clear there is a high value in the domain of crisis informatics but too often it seems to be over-invested in social media data – it feels more like after the fact analysis – more like history than designing systems that will improve the ability of first responders and persons on the ground to effectively assist in the amelioration of the disaster effects.
Also missing from the materials presented are corporate data system information and logistics intelligence – whether by default or design is not clear – which play a crucial role in the normal day-to-day functioning of affected areas such a bringing in food, services, and material goods. Companies also have a strong interest in their property, employees, and customers. Wal-Mart stores were able to function as distribution points for water and food during the immediate fallout from the Katrina crisis because of a sophisticated logistics system that at the time of the event was probably more advanced than that deployed by the federal agency responsible for crisis management – FEMA.
Instead of over-relying on tweets, posts, and text messages as one of this week’s sources noted, data scientists in crisis informatics should be thinking through the harder problems of what are the goals of the various systems affected during and after a crisis. The missing pet registry noted by Dr. Palen is clearly a case of “expanding the box” and considering the effect of 70,000 pets who also put a strain on resources and are important concerns and sources of comfort to their owners. A clearing house market to provide temporary support and reunite owners and pets is a great use of informatics systems. Alerting systems for disasters might be expanded to offer up-to-the moment resource lists presented via tweets, texts, and posts. A directory of residents based on risk factors (disabled, poor, elderly) and communication technology (land line, cell, etc) would also be a good proactive use of informatics to provide forward view guidance before the arrival of certain types of events like hurricanes. These could be cross-checked against other data sets to provide a who’s missing function for rescue groups. These can be layered with existing data maps of food stores, critical infrastructure, and shelters, and elevated land masses (in the case of flooding).
Expanding the box further: crisis informatics professionals should go beyond computer and network systems to imagine and create deployable analog solutions that can go beyond the limitations of constrained technology resources that are not available among all social groups. Drones or mini-blimps could be deployed that display electronic message boards on their surfaces to reach persons without cellular technology and to act as visual beacons to guide persons to shelter and other resources. A semantic coding based on color or other visual information (balloon shape – triangular, round, square, color – yellow, green, red) could be developed. The use of laser beacon or holographic displays should be explored. Systems of sonic information could be developed – like temporary deployable loud speakers or temporary sensitive listening devices that could hear sonic data and extract patterns of potential victim locations or other applications. What can be learned from examining the literature from theatres of war and how aid was delivered in those situations? Expanding where data comes from can help mature the crisis informatics field and differentiate it from being just a data-intensive branch of disaster sociology and social media fetish.
How can drones be deployed to drop information leaflets, solar cell chargers, water testing kits, and other useful items to isolated areas and target places where the need is greatest. And what are the policies of civilian drone use during crisis situations?
For my article this week, I chose a crisis informatics new story from the University of Pennslyvania which has developed a technique they call “social triangulation” and can be used to identify non-geotagged tweets and posts. As the lead author Rob Grace explains, “It's a very simple observation that we can look at someone's social network and infer their location based on the characteristics of that network,” said Grace. “People in a local area tend to follow local organizations. This suggests a method of some utility” (Penn State, 2017). Essentially, he’s identified a means to localize the whereabouts of social media feed without the user providing explicit geo-tagged information. While the applications for crisis informatics is beneficial how can this be used by a government to track persons of interest?
Caution should also be exercised in the exuberant wish to help and several moments of reflection be committed to how these new techniques and systems can be abused by bad actors. What can bad actors do with new techniques for extracting valuable information from social media streams? And when crisis informatics technology is deployed in politically unstable and corrupt states, can those same systems later be turned around and used to oppress and control vulnerable populations?
Takeaways: (1) crisis informatics can play an assistive role in channeling data streams into useful information that can be used by first responders and volunteers and provides a method for analyzing large patterns of behavior that can be used to identify and direct potential means coordination and response behaviors for future crisis situations. (2) “Expanding the box” – thinking beyond social media data and considering analog and other information control and communication channels would help expand the value of crisis informatics to solutions before, during and after disaster events.
What resonated with me: Over-reliance on digital technology occludes thinking of other communication & control channels and solutions that may be more robust and less dependent on existing infrastructure to be effective. Eyes in the sky that provide real-time visual/auditory/ monitoring of affected areas may open up new analytics tools that can provide assistance in finding missing persons, detecting after-effect problems (broken gas, water lines), and flow patterns of persons in a macro level analysis of behavior - just to imagine a few possibilities. Overcoming the strong bias of data scientists to work with digital forms they know and are comfortable with must give way to expanded conceptions of information streams.
Microsoft Research. (2016, August 8). Data in disasters: The potential and the limits. Retrieved from https://www.youtube.com/watch?v=-QSivtXGJS0
Palen, L. (2015, October 16). Frontiers in crisis informatics. UC Irvine's Trends in Society and Information Technology Series. Retrieved from https://www.youtube.com/watch?v=OdQRVY2tKcY
Penn State. (2017, July 20). Social media triangulation provides new approach for emergency responders. ScienceDaily. Retrieved July 27, 2017 from www.sciencedaily.com/releases/2017/07/170720095038.htm Reply Reply to Comment (1 like) Collapse SubdiscussionBrian Whitmer
Brian Whitmer Jul 28, 2017 Jul 28 at 7:28pm Manage Discussion Entry How this all ties into city government initiatives to deploy more sensor and monitoring technology made me wonder if the Department of Homeland Security and FEMA are budgeting resources to tie crisis informatics projects and research into these local government initiatives. Hopefully, there will be an actual public debate over the privacy and risk management trade-offs of deploying technology that could save lives in a disaster but could be used by bad actors to cross the traditional boundaries of privacy.
Another fascinating and sobering topic.
I hope our government officials have the wisdom to strike a good balance between safety and freedom.
Shaking the magic eight ball returns: “Reply hazy try again.”