A new system developed at the University of Rochester aims to give users genuine opinions about a restaurant before you commit your money. Called “nEmesis,” their software uses machine learning to listen to geotagged tweets that match a restaurant location. It follows that user’s tweets for 72 hours, and captures any information about them feeling ill. Though it’s not good at accounting for random bouts of the flu amidst genuine food poisonings, over a four month period it correctly identified 480 reports of food poisoning. So maybe before you risk that C-rated lunch spot in NYC, check out nEmesis to see what’s really going on.