Citizen-generated data obtained by social media listening is becoming a valuable public health tool.
Cities are notoriously noisy: The more people, activities and events condensed in a given space, the louder it is. Today, people are sharing everyday experiences at an enormous scale, and their pictures, videos, geotagged posts and check-ins create a new kind of urban noise. A new image of the city emerges from this collection of citizen-generated data and presents an opportunity to better understand urban activity. But in order to distill meaningful insights from all the noise, you need a good listener. Some government officials are beginning to catch on.
Diners who suffer food poisoning rarely report it through official channels, even though foodborne illness is a public health concern. However, sick, unhappy customers have incentive to vent their complaints on Yelp, a popular app and website for local business reviews. New York City’s Department of Health and Mental Hygiene recently completed a pilot project in partnership with the company aimed at identifying unreported outbreaks of foodborne illness. Working with software developers at Columbia University, city researchers converted nearly nine months of Yelp reviews into machine-readable data. They were then able to pinpoint potentially hazardous establishments by reviews that included terms such as “sick,” “vomit” or “food poisoning.” Scanning 294,000 restaurant reviews in New York, the software flagged three restaurants that together produced 16 documented illnesses. When health inspectors subsequently visited these establishments, they discovered astonishing health code violations: improperly sanitized surfaces and bare-hand contact with ready-to-eat food at the first two, and live roaches and evidence of mice at the third.
Food poisoning is not the only public health issue that goes severely under-reported. In order to report the side effects of prescription drugs, patients must fill out and submit a lengthy four-page form. The FDA, in partnership with Boston University and Harvard Medical School, analyzed 6.9 million Twitter posts generated over the course of seven months. In a Drug Safety paper published this April, researchers identified 4,401 tweets that described side effects worth reporting to the FDA. In fact, the complaints about gastrointestinal problems and psychiatric effects mirrored the FDA’s independent data sets on such conditions.
Aware of this critical hole in its current database, the FDA helped fund the launch of Epidemico, a health data collection and analytics startup. The company then developed MedWatcher, an app that allows people to access and easily navigate the FDA’s database, which is integrated with data from thousands of other sources, for information on drug side effects. It also provides an accessible avenue for patients to relay their experiences with drugs back to the FDA.
These examples illustrate two important points. The first is that urban airwaves are becoming increasingly proliferated with actionable public health data. According to a 2013 report from Pew Research Center’s Internet and American Life Project, city dwellers are 50 percent likelier to use Twitter than their rural and suburban peers, and increasingly geotagged social media, in the face of the rising importance of “where” in government, marks the future of hyper-local data aggregation tools. The second point is that in each example, a government agency partnered with a research university and private company. The imperative for government action to be based on data-smart strategies is gaining ground. Companies, universities and nonprofits offer a trove of data and analytical methods that government cannot afford to ignore.
Yelp and Twitter offer users intuitive social interfaces that are easy to use. They generate data because people enjoy using them. Public agencies have begun to listen in, but as cities become increasingly responsive, such third-party mediation should become unnecessary. Existing channels for communicating with the government — whether about drug side effects, food poisoning or anything else — must be improved so that instead of just listening, they can be responding in real time.