Too many big data initiatives are science projects that take months of effort, risk failure and require highly trained data scientists with scarce skills. According to a CSC survey, 55 percent of big data projects aren’t completed and many others fall short of their objectives.Read more...
Postal Service Spots Fraud with Analytics
The U.S. Postal Service continues to leverage big data analytics to spot fraudsters through an effort coordinated by its Inspector General’s office.
Dubbed RADR, short for Risk Assessment Data Repository, the big data system provides auditors and investigators with an updated overview of potential scams and other forms of mail fraud that could lead to investigations and convictions.
RADR was launched by the Postal Service as an investigative tool that compiles mountains of data to spot and investigate various forms of mail fraud. Of particular interest to postal inspectors are contract, financial and health care fraud along with mail theft.
The data analytics tool essentially provides investigators with a nation-wide threat map. The system analyzes and merges data from a variety of sources, then produces a visualization in the form of a color-coded U.S. map. The visualization helps investigators pinpoint potential hotspots and dig deeper to identify specific threats.
The RADR system was reportedly developed in-house, with the Postal Service working with a contractor to develop the algorithms needed to massage data to track down attempts at mail fraud.
The analytical tool also reportedly allows investigators to prioritize which cases they will pursue to ensue a bigger return on the time invested in each case. The amounts recovered can range as high as $1 million, officials recently told the trade publication GCN.
The RADR program was launched in 2009 and became operational in 2011. The initial focus was on fraudulent health care claims. The system was “trained” using data from different federal agencies to spot patterns like the frequency and amount of claims filed with health insurers. A clear pattern would result in a red hotspot on the nation-wide RADR threat map, the highest priority for investigators.
Medium-level risk would be depicted in the map as yellow.
The big data visualization uses geographic information systems (GIS) mapping software developed by Esri. GIS software is used to analyze and visualize data to reveal relationships, patterns and trends. Algorithms were developed for the RADR system to train the system to spot suspicious activity and help mail inspectors be more proactive in their fraud investigations.
Esri’s “geographic approach” to data analysis includes helping investigators ask the right questions, acquire the necessary data, analyze that data to spot relationships and trends and, finally, act on specific threats.
Program managers said RADR can help investigators determine which cases pose the greatest risk to the postal service so that scarce enforcement resources can be deployed to achieve the biggest return on investment in manpower and resources.