This morning I read an important guide by IBM, Making Critical Connections: Predictive Analytics in Government, Improve Strategic and tactical decision-making. The guide highlights the benefits of predictive analytics in government and shows how data can provide insight and guide decision making for government agencies. I’d encourage you to take a look at the report, and consider how predictive analytics can help transform your agency.
Today, government collects more data than ever before. There are countless examples, whether data is coming from social media, photo sharing, business operations, customer service, HR, health data, research studies, there are dozens of examples of how much data is collected. With all this data now readily available, the challenge becomes how to manage, store and drive action based on data. Predictive analytics has become a potential way to unlock data driven decisions for agencies. In discussions with government employees, here are some of the common challenges related to predictive analytics:
Just like any kind of implementation of technology, it is important to consider what exactly is trying to be solved by using predictive analytics. Any successful program has clear metrics that the organization is aiming to reach. For example, if an agency desires to reduce call center volumes, predictive analytics can be used to identify challenges and issues that customers are facing. If there is a common thread or theme seen through the data, an agency can solve the problem, and the center may receive less calls.
Predictive analytics is all about data. In order to truly capitalize on predictive analytics, agencies need to identify what data they are collecting, where it is housed, who has access, how data is currently being used, and what kind of data they may need to look at to solve the problem they identified.
Who else is Using Predictive Analytics? How can we use predictive analytics?
There are a lot of agencies using predictive analytics, and there are many great ways that agencies have implemented predictive analytics into their agency. Recently, I wrote about the Blue CRUSH program in the City of Memphis, as a great example of using predictive analytics to fight crime in our neighborhoods. GovLoop also hosted a webinar that discusses the Blue CRUSH program, which you can access here.
Another great resource comes from Frank Stein, Director of IBM’s Analytics Solution Center, also provided some great examples of using big data in his post Big Weather, Big Data.Further, the IBM guide further identifies some ways agencies have used predictive analytics, stating:
- Law enforcement agencies look for patterns in criminal behavior and suspicious activity. This enables them to deploy personnel more effectively and to identify possible motives and suspects.
- Auditors of tax returns and Medicare/Medicaid claims compare information across cases to understand normal activity patterns. In this way, they can identify cases that deviate from the norm and, therefore, warrant further investigation.
- Disease management analysts study events that led to favorable outcomes across time and patient populations, in order to develop optimal treatment protocols.
- Public health authorities monitor syndromic information from various sources, looking for elevated levels of certain symptoms that signal a widespread disease outbreak. This accelerates the process of uncovering the cause of the outbreak.
- Network analysts protect the security of computer and communications systems by detecting “cyber threats.” These include unauthorized access and the release of computer worms or viruses.
There are many reasons why agencies decide to use predictive analytics. IBM provides an extensive list of the benefits below. In today’s climate for government employees, predictive analytics can help conquer some of the most pressing challenges for government. For instance, predictive analytics has successful lead to a reduction in costs, identifying new efficiencies, and improved safety, and managing risk. IBM expands on these themes in their report:
- Reduce costs while improving resource allocation. Facing an increasing backlog of collections, an agency develops a collection prioritization plan that leverages its limited resources and aligns operations with new strategic goals. By focusing its collection efforts, the agency achieves a higher success rate, resulting in additional annual revenue.
- Reduce fraud, waste and abuse. A Medicaid fraud detection office predicts which claims are likely to be fraudulent, so that auditors can concentrate on the right claims and recoup lost revenue more cost effectively.
- More efficiently protect public safety and security. Analysts at a U.S. metropolitan police department review and analyze crime data, identify trends and patterns, and develop predictive models that are then made available to operational personnel through an intranet. Command staff can evaluate real-time conditions and send police units where they are most likely to be needed.
- Better manage risk. Government agencies are alerted to anomalies in the reported number of cases of a particular illness. As a result, medical personnel in the affected area can be notified in a timely fashion.
- Streamline processes. Millions of pieces of data from microarray experiments, such as genetic factors underlying malignant brain tumors in children, are analyzed to discover the most effective therapies, thereby extending or saving lives.
- Increase job effectiveness. Recruiters are able to improve their efficiency at filling jobs by focusing on the few candidates among hundreds of leads that are most likely to respond favorably.
The complexity of data and the challenges faced by government today require new ways of thinking, new technology and for agencies to continually learn how to provide improved services more efficiently. As pressures continue to grow on agencies to improve service delivery, while budgets and resources decrease, predictive analytics is one way agencies can find new and innovative ways to transform their agency with data drive decisions.