Big Data = Big Benefits

9/1/2014

STATE LEGISLATURES MAGAZINE | SEPTEMBER  2014

States are using large amounts of data to improve efficiency, fight fraud and identify savings.

By JoAnne Bourquard and Cassandra Kirsch

Although the term “big data” sounds vaguely sinister—like a relative of Big Brother or Big Government—it’s an unfair rap. At least in the case of state governments, it is being used to increase public safety, uncover fraud, save money, create efficiencies, and improve health and human services, among other things.

Big data is exactly what it implies: collections of data so vast that the time and effort required to manage and sort through them make traditional approaches of analysis infeasible. Yet, within these big data collections lie valuable patterns and useful information now being mined with the advent of predictive analytics software.

By cross-referencing information from a variety of sources, predictive analytics can uncover previously obscure connections between events and people and reveal anomalies and inconsistencies worth investigating. Overlaying data about health care, transportation or crime, for example, with geographical information can identify exactly where the greatest needs or problems exist, helping policymakers allocate resources most effectively. 

For several years, businesses have been collecting enormous amounts of data about consumers and their lifestyles from mobile and web technologies, which they use to track customers’ habits and tailor marketing materials. “Your smartphone is not just a phone. It’s a tracking device. ... And if I can tap into that information, I can predict and persuade your behavior by targeting it with the right ads,” says Hewlett Packard’s Chris Surdak, author of “Data Crush: How the Information Tidal Wave is Driving New Business Opportunities.”

Now state governments—generators of large amounts data themselves—have begun tapping into the wealth of possibilities big data offers to transform operations, such as improving public safety and health care, unearthing fraud and much more.

Of all the data now captured, experts believe about a third of it could be useful, yet only 0.5 percent of it is currently being mined. Richard Leadbeater, global solutions manager at Esri, a firm that specializes in geographic information system technology, says that big data has such significant economic potential that it should be viewed as a natural resource.

A recent report by the National Association of State Chief Information Officers describes state governments as “enormous data generation engines.” State officials who treat this data as an asset, analyzing it to discover new patterns, correlations and insights, the report says, can “gain a competitive advantage.” Here are some ways states are doing just that in various areas.

Tax Fraud

Big data is helping uncover tax fraud which cost states billions of dollars annually. Connecticut, Georgia, Indiana, Louisiana, New York, South Carolina and Washington, among other states, have contracted with big data companies to use data analytics technology to detect tax schemes and reduce the number of fraudulent refunds states pay out each tax season. Using identity-based filters, the technology screens the information contained in each tax refund request against billions of personal records from public and commercial databases to snag the bad actors.

Georgia has caught $25 million in fraudulent refunds since it began using big data in 2012. Washington’s program has begun recouping more than $10 million a year, and New York’s Department of Taxation and Finance decreased the revenue drain caused by fraudulent refunds by $1.2 billion in 2010 alone, in addition to increasing tax collections by more than $100 million.

Public Safety

Police departments across the country have begun contracting with big data companies to help detect and predict crime patterns. In New York, Las Vegas, Los Angeles and Memphis, data are being mined from arrest records, surveillance video and social media.  In Las Vegas, police officers use past data sets to predict which locations are likely to be hit by future crimes, so they can target their limited resources carefully.

North Carolina lawmakers’ attention turned to big data after two university students were murdered by a couple of men who would have been in custody at the time of the crime if court officials had been able to access the data on current offenders. Recognizing the benefit to state and local law enforcement officers and court officials of being able to quickly obtain that kind of information, the General Assembly allocated funding to develop a program for sharing data among these agencies.

The system that was developed, called the Criminal Justice Law Enforcement Automated Data Service (CJLEADS), integrates information from various databases—warrants, jail records, court records, prison records, probation and parole status, sex offender registration, motor vehicle, wildlife and concealed handgun permits—and delivers a comprehensive, real-time picture of individuals quickly via the Web.

“If law enforcement can save five minutes each time they query CJLEADS, versus logging onto multiple systems, the value of that time savings is an estimated $18.6 million annually,” says Kay Meyer, former project director in the office of the state controller who led the effort to design and implement the system.

“Put simply, CJLEADS improves public safety. If a single life is saved because of the system, its value is truly priceless,” says Meyer, who currently is an industry consultant for SAS, the North Carolina-based software analytics firm that developed and hosts CJLEADS.

In Alabama, the Mobile Officer Virtual Environment program also allows officers to pull real-time data on potential criminals.

Health Care

To contain costs and better target limited Medicaid dollars, states are beginning to use data in new ways to reduce costly, and sometimes preventable, hospital readmissions. Officials in Maryland, Kentucky and Washington, for example, are using data from their health information exchanges to analyze hospital admissions and readmissions.

One pattern they are finding is that patients who use hospital emergency departments often usually need other services but do not know how or where to find them. According to the Kentucky Department of Medicaid, 79 percent of the patients who visited the emergency room 10 or more times in a 12-month period had behavioral health issues, and 45 percent had been diagnosed with a substance abuse problem. Using big data, doctors can now direct many of these patients to clinics or other less- expensive care centers, track their conditions and provide more appropriate care.

Human Services

Los Angeles County faced increasing threats of fraud in its child care program for low-income families. Much of the fraud was perpetrated by organized fraud rings that recruited parents as well as child care providers, making investigations complicated and difficult to prosecute.

To combat fraud and ensure that benefits would be available to families truly in need, the Los Angeles County Department of Public Social Services used a data mining technique that identifies and expedites the review of suspicious cases. The pilot project achieved an 85 percent accuracy rate in 2008-09 in identifying public assistance fraud rings and saved an estimated $6.8 million in avoided losses due to fraud. 

Using this same technique, between May 2011 and July 2013, the Los Angeles child care program referred 28 cases to the district attorney for felony prosecution and initiated 405 fraud referrals. In addition, caseworkers made 753 referrals, resulting in denial, reduction or termination of benefits, and addressed overpayments as well as other fraud.

Based on current estimates, Michael Sylvester with the Los Angeles County Department of Public Social Services projects the data mining solution could save tens of millions of dollars over the life of the project, although any actual savings and efficiencies have not yet been determined, he cautions.

Elsewhere, Indiana recently launched an initiative to use big data analysis to lower its infant mortality and child fatality rates. The state’s new Management and Performance Hub will provide centralized data sharing and analysis across several agencies.

One of the first projects will be to analyze data from various databases, including criminal history, health and demographics, to gain a better understanding of the causes of infant and child deaths and to identify potential solutions. The long-term goal of the system is to share data among all the executive branch systems and use analytics to improve state decision making in a variety of areas.

Role of the Legislature

Legislators are playing a key role in the growth of big data centers within their borders. To encourage their development, in 2013, lawmakers in Arizona and Illinois passed bills to exempt data center equipment from sales taxes, and Texas lawmakers chose to eliminate most of the sales taxes for new data centers that are larger than 100,000 square feet and require at least a $200 million investment.

Legislatures in Colorado, Florida, Massachusetts, Nebraska, and New Jersey have considered legislation this year related to big data. A new Colorado law allows tax exemptions for the development of qualified data centers. New Jersey’s new law directs the state’s Big Data Alliance to encourage state government, academia and industry to study and suggest solutions for the significant and immediate challenges posed by the proliferation of big data sources and the resultant deluge of digital data.

“Big data isn’t just some buzzword. Industry, academia and government have touched only the tip of the iceberg with big data,” says Assemblyman Upendra J. Chivukula (D), deputy speaker and chair of the Assembly Telecommunications and Utilities Committee. “Big data poses both challenges and opportunities. By committing to developing a strategic approach and coupling that with new advanced cyber-infrastructure, we are positioning New Jersey as a place where innovation can thrive.”

Massachusetts Senator Gale Candaras (D), chair of the Joint Committee on Economic Development and Emerging Technologies, sponsored legislation that would create the Big Data Innovation and Workforce fund to help prepare the state’s future employees for careers in big data. The fund also would ensure that academia and industry have the necessary tools to use the data to identify and solve problems in transportation, public health and energy, as well as support economic development.

She emphasized the fund would help “to close the skills gap” confronting a future big data workforce and provide “challenge grants for the Commonwealth’s agencies, departments and municipalities to use with big data to solve public policy concerns.”

Florida lawmakers passed legislation this year that consolidates several state agency data centers and computing facilities into a single state data center. And Nebraska senators passed a state policy declaration to “make revisions in Nebraska’s tax structure in order to ... promote the creation and retention of new, quality jobs in Nebraska, specifically jobs related to ... large data centers.”

Analytics Ahead

Data analytics is revolutionizing how states access and analyze information, improving not only the bottom line, but also how state agencies predict and meet constituents’ future needs.

Data Analysis Uncovers Fraud in North Carolina

The success of North Carolina’s Criminal Justice Law Enforcement Automated Data Service led to the creation of the North Carolina Government Data Analytics Center, which allows state agencies to share data. One of its goals has been better data analysis to identify fraud, waste and compliance in areas such as unemployment insurance and workers’ compensation.

“One of the first projects we tackled was unemployment insurance, as that program had nearly $3 billion in debt,” says North Carolina Senator Ralph Hise (R), chair of the Senate Health Committee and co-chair of the Senate Appropriations on Health and Human Services. “IT systems at the Employment Security Commission, labor department and tax office couldn’t talk to one another, which made it difficult to identify fraudulent claims. As a result of data retrieved from the center and other changes to governing rules and regulations, we reduced that debt to under $1 billion in 16 months.”

The center’s success in identifying fraud and abuse related to unemployment insurance garnered enough support from lawmakers to expand it to more areas. “We want to be able to answer questions using data to evaluate the effectiveness of state government programs in a variety of areas, and GDAC makes that possible,” says Hise.

North Carolina Representative Jason Saine (R), who chairs the House Appropriations Subcommittee on Information Technology and co-chairs the Joint Legislative Oversight Committee on Information Technology adds, “GDAC is transforming what we do and how we make decisions. Information technology touches everything we do in state government, and its importance is only going to grow.”

BIG DATA

By the Numbers

2.8 Zetabytes

The amount of information stored worldwide in 2012 

140 Zetabytes

The amount of information estimated to be stored worldwide in 2020  

0.5%

Percentage of all the world’s data currently analyzed

33%

Percentage of existing data in the world that could be useful if properly tagged and analyzed

Over 5,000 miles

The estimated cumulative size of all the world’s data centers in 2016

Sources: SAS, a North Carolina-based software analytics firm, 2014

 

 

Leveraging IT Partners Project

NCSL’s Leveraging IT to Strengthen Government Partners Project, sponsored by the NCSL Foundation for State Legislatures and private sector partners, provides information about big data, data analytics, location analysis and more.

The project is exploring ways to strengthen government decision making through improved business processes, effective data management and the use of advanced analytics to identify and combat fraud and abuse.

JoAnne Bourquard is a senior fellow at NCSL who covers technology issues. Cassandra Kirsch also tracks technology issues for NCSL.

 

Additional Resources