Found a really cool article on how the city of Chicago teamed up with the Civic Consulting Alliance and researchers from Allstate Insurance’s Quantitative Research & Analytics Department to put together a predictive model using 311 calls, previous inspections, permits, and other data points to identify high risk restaurants. Here is what I found most interesting:
- Leverages public data to identify Chicago restaurants most likely to face health code challenges, so health inspectors to prioritize inspections for those restaurants.
- Using 311 requests, sanitation complaints at restaurants and information on previous inspections and permits, the researchers identified a group of factors that could lead to health code violations.
- These factors were then brought together to create a predictive analytics model that was used to identify which restaurants should be inspected first in order to stop critical violations that could lead to illnesses.
- The model evaluation calculates risk scores for more than 10,000 Chicago food establishments using publically available data, most of which is updated nightly on Chicago’s data portal.
Using a tool like OpsAnalitica you can generate reports and conduct analysis like the city of Chicago on your locations. They aren’t doing anything super special they are just using data points that they collect from a couple of different systems and bringing it together to identify risk factors. If you were doing daily line checks, temp logs, daily logs and combining that data with sales, costs, and customer service data; you could have a complete understanding of how your operations really work and what things are driving costs and reducing profits.
I have clipped the article below, or to see the original article click here.
Chicago is using predictive analytics to better ensure food safety for city residents and visitors.
The new system, built by the Chicago Department of Public Health (CDPH) and Department of Innovation and Technology, leverages public data to identify Chicago restaurants most likely to face health code challenges, so health inspectors to prioritize inspections for those restaurants. “The use of open data will result in a more streamlined approach to overseeing food safety, targeting our resources at higher-risk establishments without compromising safety oversight at any food business across the city,” Mayor Rahm Emanuel said.
Research for the new system was conducted in collaboration with Civic Consulting Alliance and researchers from Allstate Insurance’s Quantitative Research & Analytics Department to identify various factors that can result in a restaurant facing health code violations.
Using 311 requests, sanitation complaints at restaurants and information on previous inspections and permits, the researchers identified a group of factors that could lead to health code violations. These factors were then brought together to create a predictive analytics model that was used to identify which restaurants should be inspected first in order to stop critical violations that could lead to illnesses.
The model evaluation calculates risk scores for more than 10,000 Chicago food establishments using publically available data, most of which is updated nightly on Chicago’s data portal.
CDPH performed a simulation that allowed it to identify establishments with more violations using the predictive model than using risk categories alone. As a result of these findings, CDPH is now using the predictive model to prioritize inspections among the highest risk establishments.
“Researchers, cities and the public are able to freely access and download all the information needed to replicate the research,” said Brenna Berman, CIO of the Department of Innovation and Technology. “Releasing the data and research on GitHub allows for collaboration with other scientists and institutions to improve the city’s forecasts and allows the technique to be adopted by other cities conducting food inspections.”
Funding for the project came from a $1 million award in 2013 from the Bloomberg Philanthropies’ Mayors Challenge to develop a data-based approach to decision making.
This isn’t the first time Chicago has used analytics to protect residents from food-borne illness. In 2013, the city developed a predictive analytics model to fight its growing rat population. The Department of Innovation and Technology used 311 city requests ranging from stray animal calls to vacant and abandoned buildings as well as missing or overflowing garbage cans and restaurant code violations. Through the requests and the model built to analyze them, the city could visualize on a map where and when rat populations spiked.
The same year, researchers in Chicago created another program called Foodborne Chicago, which examined Twitter for food poisoning complaints from residents, employees and tourists in the city. The program led to 150 additional inspections by the FSD that year.
