Data-Smart City Solutions


By Nick Carney

Much of the attention lavished on Twitter lately has revolved around the site’s initial public offering and diagnosing what the event tells us about the knowledge economy. And of course, there’s always the entertainment of a good celebrity meltdown to reinforce the central role that the tweeting plays in contemporary pop culture.

But the apparent genius of Twitter lies not just in its profitability as a business or its status as the modern-day soapbox for drunken diatribes issued in 140-character intervals; public policy stands to gain from the rise of the twitterati as well, particularly health policy.

While we have previously discussed the opportunities presented by curating public sentiment from social media like Twitter or Yelp, researchers are now beginning to wield Twitter as a tool to fight the outbreak of disease.

Some of the more innovative work on a city level deals with preventing outbreaks of foodborne illness that result from unhealthy restaurants.

One strategy in this war against food poisoning involves employing social media to encourage residents to participate in the existing, formalized food poisoning tracking system. The City of Chicago, for instance, has taken to Twitter to directly contact those residents who complain of foodborne illness. The scheme, named Foodborne Chicago, uses a program designed to detect language that suggests an individual is suffering from food poisoning. After identifying local Twitter users who allege that Windy City restaurants caused a bout of food poisoning, the City uses volunteers to contact the afflicted resident and suggest that he or she file a complaint with the Department of Public Health. Created by the Smart Chicago Collaborative, a quasi-public civic organization, as of August 2013 the program had inspired 33 restaurant inspections as a result of having contacted 150 Chicagoan Tweeters.

The other route is to involve users only passively. Scientists at the University of Rochester chose this route, and developed an algorithm that sweeps the Twitterverse for evidence of food-borne illness at particular restaurants. Named nEmisis, the tool used GPS data on the location of the 29,904 restaurants inspected by the City of New York and tracked geocoded tweets sent from those locations. After detecting a tweet sent from a restaurant, the software continued to monitor tweets from that account. An automated language model developed by the authors determined whether the individual tweeted about health effects symptomatic of foodborne illness in the 100 hours following the restaurant visit. The results correlated well with official data on outbreaks of food poisoning, indicating the tool’s reliability and the potential to improve existing official data.

Work has also begun on using the public musing of Twitteranians to tackle even more serious diseases. Researchers from Cornell and Carnegie Mellon recently demonstrated that Twitter data could be employed to track influenza outbreaks as well. The authors produce an algorithm that can sift through geocoded tweets containing various flu-related indicator words and produce a real-time model of influenza outbreaks. Furthermore, the data provides clear information on the phases of a flu outbreak, from the non-epidemic phase to epidemic phase, allowing for a nuanced account of how quickly the disease is spreading.  They found that the Twitter data was highly correlated with flu data obtained from the Center for Disease Control, and that present evidence from their Twitter-based model more accurately reflected the disease outbreak than a model based on other digital data sources, such as Google searches.

In short, Twitter appears to be a potent weapon for public health organizations working to accurately predict disease outbreaks in their early stages, allowing faster response times and more efficient containment of diseases.

So no more is Twitter just for drunken exposition or the proliferation of interesting internet links; that annoying guy who always tweets about how he’s feeling in the morning may one day unwittingly contribute to an effort to save your life. Never has the health of so many depended so on the oversharing habits of so few.

About the Author

Nick Carney

Nick Carney is a Masters in Public Policy candidate at the Harvard Kennedy School of Government and a Public Service Fellow. He is concentrating in Social and Urban Policy and has previously worked in clean energy policy and mixed-use, urban real estate development. Nick is particularly fascinated by transportation, land-use, and education policy, and with his sister runs a literacy nonprofit called Breaking the Chain.



comments powered by Disqus