By Jeremy Clopton, CFE, CPA, ACDA
As an investigator, much of the data I analyze on a day-to-day basis is standard financial data – accounts payable, payroll and general ledgers. For those of you using data analytics, you likely see these data sets as well. However, a recent presentation by a colleague of mine and an article on insurance fraud reminded me of the many other data sets that are useful in investigations.
My colleague and I work closely together, but we do not always have the chance to discuss the behind-the-scenes details of what we do. Always looking for a new way to apply data analytics, I was definitely intrigued by his presentation on predictive coding and analytics surrounding “unstructured” data like email. I quickly realized that email, much like financial data, is just as ripe for mining as accounts payable data. It takes some prepping and cleansing, but at the end of the day, it contains the same types of information: names, addresses, dates, times and keywords. Considering these documents as data can open your eyes to a completely new set of potential procedures in your next investigation.
This recent article on insurance fraud did not specifically cover data analytics, though it did highlight some potential applications and red flags. Applying data analytics to claims and employee data can help identify many of these red flags (new employee, poor attendance, late Friday/early Monday claims, unusual location, etc.). For example, an attendee at a recent presentation I gave talked to me about using visual analytics to map insurance claims and identify “hot spots” in a city. Plotting all claim addresses on a map allowed them to focus on streets and neighborhoods with multiple claims during a specific period. This visual analytic helped them discover fraudulent claims.
Social Media Analytics
The beginning of the article also described how an individual who filed for workers’ compensation got busted after appearing on the game show, “The Price is Right.” I’m not advocating watching hours of game shows to identify potential workers’ compensation fraud; however, analyzing social media data could be just as helpful. With Twitter, Facebook, Instagram, Pinterest, LinkedIn and the many other sites out there, individuals create enormous amounts of data every day (much of it about what they are doing). Analyzing this data for keywords, events, times and locations can effectively supplement an investigation. Again, once obtained, this information is much like the financial data we are already accustomed to analyzing.
You may be aware of these data sources and their applicability, but if you aren’t, I hope this article has you thinking of new ways to incorporate analytics into your next investigation.