IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Chicago Fights Food Poisoning with Predictive Analytics

Chicago has discovered a way to use restaurant and neighborhood data as red flags for food inspections.

For years, cities have relied on official food inspections as their primary defense against shady restaurants and questionable dishes. But the city of Chicago is raising the stakes through predictive analytics, which not only helps officials catalog offenders, but also identifies high-risk restaurants for citizens and inspectors.

Chicago CIO Brenna Berman and Chief Data Officer Tom Schenk completed a pilot program this month that takes data sets from the more than 15,000 restaurants in Chicago and its surrounding neighborhoods, and estimates which eatery is most susceptible to health issues. Data scientists from AllState Insurance contributed to the project, and data was gathered from the city’s WindyGrid system, a data repository that collects real-time information from city departments at a rate of 7 million rows per day.

Berman’s team, operating within the city’s SmartData analytics platform, harnessed a number of data types to make predictions. Some data sets include information specific to restaurants, such as the age of the business and previous inspection scores. Yet the analytics pilot also factored in corollary data from sanitation complaints and the occurrence of property-based crimes.

“What we’re trying to answer here is, on any given day, what are the restaurants that are going to be the biggest priority?” Schenk said.

The ultimate vision for the analytics project is for the city’s Department of Public Health to dispatch inspectors based on need versus random appointment. Additionally, Chicago's health inspectors will eventually be dispatched to restaurants by ZIP code or restaurant types for convenience. The city is documenting granular details of its pilot so other municipalities can replicate it locally.

“It’s about providing a broader view of analytics, not just taking a look at data, but also taking a look at predictions and putting that into the hands of users,” said Schenk .

While more pilots are pending, teams at Chicago’s Department of Innovation and Technology, led by Berman, are dedicating a major effort to develop WindyGrid as a free open source platform so other cities can launch their own data initiatives. It will be a software easily downloaded and manipulated on open source platforms like Github.

Once finished, Berman estimated WindyGrid's development to total roughly $3 million, $1 million of which has come from a grant from Bloomberg Philanthropies Mayor’s Challenge, a contest awarding innovative endeavors proposed by cities, and the other funding generated by the city. Despite costs, Berman said the WindyGrid’s value to Chicago — and other cities — far outweighs the expenditures.

“We finished this really important predictive analytics use case, and launched a couple of others, but are marching down the path of our commitment to open source,” Berman said.

The goal is to release WindyGrid’s updated code via Github and other outlets sometime in the fall — the ultimate test for the release being outside adoption.

“The true measure of success is only going to be found after we get through this development effort,” Schenk said.

Jason Shueh is a former staff writer for Government Technology magazine.