Making Predictive Analytics Scale: SmartData Chicago

October 27, 2014 7:57 am PST | Data Apps & Visualization, Data as a Service

Analytics doesn’t scale. And that’s a problem for Chicago (and many other governments.) The city has one dedicated team conducting advanced analytics for all 33 of its departments. And while that team is certainly helping the Windy City make more sense of its data and driving it to better decision making and greater efficiency, it’s just not enough.

“The issue is that analytics only scales linearly,” says Tom Schenk, Jr., Director of Analytics and Performance for the City of Chicago, “So, right now, if we want conduct 50% more advanced research projects, we need 50% more staff.”

Data-based decision making is conducted at a department level in Chicago, but, for advanced analytics, agencies have to work with Schenk’s team. That causes complications like bottlenecked workflow and other limitations created when working with an outside team. To improve the volume and quality of data-based decisions, Chicago’s Department of Innovation and Technology, Event and Pattern Detection Laboratory at Carnegie Mellon University’s Heinz College, and the University of Chicago’s Computational Institute are developing the SmartData platform. The project was made possible when the City of Chicago was awarded a grant from the Bloomberg Philanthropies’ Mayors Challenge.

When it’s launched, in about a year, the SmartData Chicago platform will to decentralize analytics, allowing each department to conduct its own advanced analytics without having to understand how to compile them or contact Schenk’s team. While SmartData is in development, components are going live on a piecemeal basis. But, even in its current stage, the platform is showing signs of greatly improved efficiency with no loss of efficacy.

Take, for instance, rats. Right now there is a staffer in the Department of Sanitation who spends one to two days a week predicting where rodents may appear throughout the City of Chicago. Based on those predictions, Sanitation teams are deployed to those areas.

The SmartData platform can make determinations about where rat populations may occur in almost no time at all, and with the same accuracy that the Sanitation employee can. In the end, the platform will allow her to regain 20% of her work week and focus it on other tasks the department needs. In addition, Sanitation will less heavily have to rely on institutional knowledge of its staff.

“The SmartData platform is a tool which will increase efficiency, allowing us to get the same results in less time,” Schenk told Socrata in a recent phone interview, “It will also allow us to better deliver city services, taking care of issues before people even complain, and overall improve resident satisfaction throughout Chicago.”

Eventually, the city plans to roll out the platform to other areas such as restaurants inspections, the cigarette black market, and social services such as homelessness and abandoned buildings, according to Schenk.

If this decentralized approach to data driven government decision making sounds like something that could benefit your city, state, or agency, there’s good news. “We realize the solutions we are building for Chicago are applicable to other cities, too,” Schenk says, “So we strongly believe that whatever we develop should be open sourced for other cities and organizations.”



Previous Article
Data as a Service
How To Ignite Your Local Civic Tech Ecosystem: Derek Eder

October 29, 2014

Next Article
Data as a Service
Civic Awesome: Open Data In The News For October 24, 2014

October 24, 2014