Twitter is pretty powerful – it can tell us about breaking news, how people are reacting to a specific event, and what and where they’re eating for dinner. Twitter can even tell us where people are getting sick, specifically from food poisoning.
Twitter is pretty powerful – it can tell us about breaking news, how people are reacting to a specific event, and what and where they’re eating for dinner. Twitter can even tell us where people are getting sick, specifically from food poisoning.
A new study reports on nEmesis, a learning machine built by a Google data scientist that can track where people tweet about food poisoning.
In the study, the machine flagged relevant stomach- and food-related updates from a pool of 3.8 million tweets posted in a four-month period in New York City. Human eyes then determined what was thought to be the 6,000 most relevant tweets that seemed to indicate food poisoning so that nEmesis could learn what data to look for.
Geo-located tweets that contained phrases like “throw up,” “Pepto-Bismol” or “my tummy hurts” were flagged as being related to food-borne illness. These tweets were then data-mined and had metadata added to indicate open restaurants near the place where the user tweeted.
The health scores nEmesis assigned to restaurants based on the number of tweeters who fell ill after visiting came close to the scores food inspectors had submitted to the city’s health department.
This discovery shows the potential of Twitter’s big data for improving healthcare.