Government is constantly looking for new answers to old problems⎯answers enabled by the constant improvement in IT.

Government customers are no longer content to aim their business analytics tools solely at past events to see what might have happened, whether a security breach, a missed budget or project milestone, or some internal attempt at mischief.

In other words, they want traditional business intelligence tools to take on predictive qualities to head off bad outcomes or enhance the likelihood of projects and processes staying in band. And they don’t want this capability, called predictive analytics (PA), to belong solely to the lab or the executive suite. Its real value is deploying it at the operational level so people like caseworkers, financial analysts, or network security operators can use it.

For example, a feature written into software tools on Open Source R, a statistics language originally developed for scientists, allows analysts in a variety of fields to visualize and spot trends in data. Also, a graphical user interface can be used for building reports and dashboard applications. No programming is needed. This approach overcomes traditional barriers to PA, namely cost, complexity and the disconnect between PA models and PA use for improving mission delivery.

Agencies at the state, local and federal levels are using this technology with predictive analytics. For example, the U.S. Postal Service follows money order purchases to discern patterns that might indicate money laundering. Analysts and postal inspectors can look for repeated purchases in the same location for amounts at or just below the legal level. In another agency, financial analysts apply PA to payment and collection services where transactions number in the millions. It helps officials know in advance where they’ll have to deploy extra manpower or other resources to improve financial operations.

At the local level, a North Carolina police department with 1,700 officers developed predictive crime models deployed in mobile computer in squad cars. By knowing in advance where crime is likely to occur, the department has seen an 11 percent drop in crime.

The Louisiana Department of Children and Family Services replaced Excel spreadsheets with WebFocus predictive analytics tools to cut fraud in food stamps. By building in a geographic component to the analysis application, officials could trace the misuse of SNAP benefits to certain big-box stores and get there ahead of further transactions.

As I see it, predictive analytics of this sort can help a lot with government’s recurrent challenges like performance management, maximizing revenue, reducing improper payments and improving service.

Steve Charles is co-founder and executive vice president of immixGroup, Inc., which helps technology companies do business with the government. He is also a member of Breaking Gov’s Editorial Advisory Council.