System helps detect foodborne illness through Twitter

CorporateWoman_ChecksPhone_BigstockM_blogAs the old adage goes, you are what you eat, but in this case it’s what you tweet that may be more important.

Researchers at the University of Rochester have developed a system called nEmesis that analyzes millions of tweets to zero-in on people posting about foodborne illness after visiting a restaurant. The system not only can help patrons make educated decisions on where to eat but it also can help complement traditional methods for tracking food safety and foodborne illness outbreaks, according to a statement from the University of Rochester.

During a four month period, nEmesis analyzed 3.8 million tweets from more than 94,000 unique users in New York City. Of those, it found 23,000 restaurant patrons and 480 suspected or likely cases of food poisoning. The system works by monitoring public tweets and separating out restaurant visits by matching the location of the tweet with the location of the restaurant. nEmesis then follows that person’s tweets for the next 72 hours. If the user tweets about being sick (e.g.,  symptoms such as vomiting or diarrhea), the system now has that information along with information about the restaurant the person visited.

Although the reports from Twitter aren’t a perfect indicator, the data collected from the 140-character or less tweets correlates well with the New York City’s Department of Public Health and Mental Hygiene records of restaurant inspections.

“The Twitter reports are not an exact indicator – any individual case could well be due to factors unrelated to the restaurant meal – but in aggregate the numbers are revealing,” said Henry Kautz, chair on the computer science department at the University of Rochester and co-author on the paper about nEmesis, in a statement. Kautz goes on to say that a “seemingly random collection on online rants becomes an actionable alert.”

Twitter has more than 550 million active users with almost 60 million tweets posted daily, according to Statistic Brain.

Comments are closed.