nEmesis has been developed by researchers at New York’s University of Rochester. The software uses GPS data (the ‘geotags’ present when a tweet is sent from a specific location) and public tweets to try and link symptoms of foodborne illnesses with potentially responsible eateries. The tweets of a user are monitored for three days after the person has eaten in a restaurant in order to see if they subsequently show signs of flu or sickness.

A four month long trial period - involving 94,000 unique Twitter users and 23,000 restaurant visitors - identified almost 500 cases of potential food poisoning. Initial reports by the research team suggest that the data matches up “fairly well” with confirmed reports from hygiene and other public inspectors.

The team hope the data will be used to develop a heavily automated system (combined with official reports) that will eventually be able to identify what restaurants are more likely to result in cases of food poisoning. Hungry customers could then easily check out the data before settling on any given dining location. One theoretical use would be to create a faster and more robust warning system for suspected cases of food poisoning.

The Rochester researchers developed nEmesis after using ‘crowd-sourced’ human workers through Amazon’s Mechanical Tusk. These workers helped manually identify signs and trends for the software to search for.

There is some work to do before nEmesis is ready for public release. Restaurant owners will undoubtedly be concerned that an automated programme may incorrectly identify their business as a source of food poisoning, while the system will rely on a huge flow of data to identify trends and create aggregated reports. Researcher Henry Kautz admits “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: seemingly random collection of online rants becomes an actionable alert."

Given the various online privacy controversies recently, many Twitter users may also be reluctant to allow another automated system to scan their social networking feeds.