You know the saying "an ounce of prevention is worth a pound of cure." From the state department of revenue to the regional water authority, the need for preventive measures is becoming a key driver of predictive analytics' rapid growth in the public sector.
Prevention through Prediction
How does predictive analytics support the premise of prevention, and what do we really mean by prevention as it applies to predictive analytics? In this case, prevention simply means identifying a problem from data or noticing the inclination for a problem before it grows into a much larger issue. Based on a series of proven algorithms, these what-if scenarios provide the user with insight into future probabilities and trends.
The user of today is no longer the high-level statistician or computer scientist from the early days of predictive analytics. Now, if a department of revenue wants to mine its database of past tax returns to build a model to predict what the anticipated audit recovery would be for a return, it can. While the behind-the-scenes technology and algorithms are relatively complex, the interface that surrounds the algorithms should be easy to learn. No longer does one need a Ph.D. in statistics to run this type of technology. It all comes down to building the model; then anyone can use that model to take pre-emptive action.
In government circles, the rapidly growing trend in predictive analytics is a shift in its use by only a few specialist statisticians to a far larger number of government officials. The growing expertise of front-line government officials in harnessing predictive analytics as a tool to help their everyday work is driving the trend toward adoption of the software in the public sector.
It stands to reason that auditors, physicians, engineers, career military and law enforcers know what types of data they have available, and knowing how to ask the right questions of that data when looking for answers will help them make more accurate decisions.
For example, if a customs agent is trying to determine what cargo containers should be inspected, it would be helpful to use predictive analytics to first build a set of criteria to help spot any containers that, for a variety of reasons, do not match up with the norm.
Most adopters of predictive analytics are not statisticians or analysts; rather, they are experts in specific government domains (such as auditors, physicians, retired nurses). With minimum training to launch and deploy the often easy-to-understand predictive analytics graphical user interface (GUI), the government is placing powerful analytic tools in the hands of people who know their jobs inside and out.
There are several key things to look for in a predictive analytics tool:
- A broad range of capabilities, from data preparation to visualization and analysis techniques;
- An intuitive interface; and
- A range of options in deploying (putting to use) the completed analysis.
Following are some recent examples of how predictive analytics is being used at various levels of government.
The government is expanding fraud detection activities within the Medicaid and Medicare programs. Given the potential for fraud and abuse, states are now investigating claims for patterns that might be indicative of fraudulent providers. For example, the New York Department of Health was recently concerned with providers who were billing to shared or stolen Medicaid recipient identifications. By using predictive analytics, it found thousands of questionable patterns residing within the data — patterns that resulted in the department identifying false claims and fraudulent providers.
Not only do all states face compliance regulations when it comes to taxes, but they all face challenges in collecting every cent of tax due. Commonly referred to as the tax gap, it is the area between the tax owed and amount collected, which is the thrust behind tax audits: to reclaim that missing money.
The state of Texas decided to use analytics to build predictive models for both audit selection and identification of tax non-filers. As with many states, it had substantial volumes of data that were not being used properly in reducing the tax gap. As a result of using predictive analytics, the state of Texas has recovered $400 million in unpaid taxes since implementation of the software more than five years ago.
U.S. Department of Defense
In the military, recruitment is top of mind today more than ever. In an effort to increase the efficiency of recruiting and the quality of enlisted recruits, the U.S. Department of Defense has turned to predictive analytics. For example, Army and Navy officials are now building models designed to predict a potential recruit's success based on such factors as whether or not a candidate was the captain of his football team, in the glee club or was the treasurer of her senior class. The same approach and methodology goes into retention models that might serve to increase the percentage of soldiers, sailors, Marines and airmen deciding to reenlist. For example, fitness reports and soldiers' responses to satisfaction surveys are types of data used to build retention models. Such models can identify attitudes and behaviors indicative of a nonreenlistment decision, and the Department of Defense can take pre-emptive action.
The Department of Defense is also using predictive analytics for risk assessment and combat readiness of troops at home and abroad. Risk assessment is very similar to force deployment in law enforcement — the ability to predict with great accuracy the time and place of potentially harmful or dangerous activity.
Law enforcement agencies across the country face a daunting daily task - deciding where to most effectively deploy resources. In order to help make the most effective decisions, predictive analytics can exploit a variety of data sources to include calls for service, incident data that records type/location/
suspects/weapons, weather patterns and scheduled city events. The resultant models can be used to forecast locations that merit increased police presence in a particular time period based on past activity in those locations, predicted weather patterns and scheduled events during that period. Precinct captains can make informed decisions on force deployment and where best to put the department's resources at any given time.
A Bright Future
These are just a handful of examples demonstrating the range of capabilities predictive analytics can provide for different governmental entities. It allows public safety and other government agencies to anticipate behavior and events and, by doing so, maximize their resources. Whether an organization is local or national, involved in public health, defense, law enforcement, intelligence, tax and revenue or other services, the technology can be used to analyze large amounts of data, discover patterns and irregularities, conduct program evaluations and guide decision-making at all levels.
The everyday use of predictive analytics will continue to increase in the public sector because, although complex at its core, on the surface the technology is easy to use and manage. This key factor allows the tools to be in the hands of those who know their jobs better than anyone else and need the technology the most. Moreover, with so much information at hand, more government sectors will discover new opportunities to implement the technology within their own areas.
Predictive analytics enables proactive risk management, refining key decision-making processes through controlled, iterative testing of potential actions and their likely intended - and unintended - consequences. This unique insight has already helped countless government departments and agencies over the past 10 years, and predictive analytics' adoption will only increase as officials find new and innovative ways to utilize this powerful technology.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access