Welcome


Welcome to Understanding Link Analysis. The purpose of my site is to discuss the methods behind leveraging visual analytics to discover answers and patterns buried within data sets.

Visual analytics provides a proactive response to threats and risks by holistically examining information. As opposed to traditional data mining, by visualizing information, patterns of activity that run contrary to normal activity surface within very few occurances.

We can dive into thousands of insurance fraud claims to discover clusters of interrelated parties involved in a staged accident ring.

We can examine months of burglary reports to find a pattern leading back to a suspect.

With the new generation of visualization software our team is developing, we can dive into massive data sets and visually find new trends, patterns and threats that would take hours or days using conventional data mining.

The eye processes information much more rapidly when information is presented as images, this has been true since children started learning to read. As our instinct develops over time so does our ability to process complex concepts through visual identification. This is the power of visual analysis that I focus on in my site.

All information and data used in articles on this site is randomly generated with no relation to actual individuals or companies.

Using Geospatial Analysis To Investigate Insurance Fraud

Many of my articles have covered the importance of visual analysis to understanding the relationships between entities in data. Visual analysis is also important in understanding the relevance of the location in where events happens to determine and discover patterns in insurance fraud behavior.

There are numerous ways we can leverage geospatial analysis to analyze and discover potentially fraudulent insurance claims. Geospatial analysis can also be key in developing investigative plans for ring activity. In this article we are going to cover the difference insurance fraud scenarios that geospatial analysis plays a key role in the strategic and tactical analytical discovery of insurance fraud.

Organized Ring Activity

In previous articles we have reviewed how to use link or association analysis to pro-actively discover organized fraud activity such as staged accident rings and medical provider fraud. Now we are going to incorporate geospatial analysis to visualize the location of the activity for the development of an investigative plan.

As an insurance fraud analyst, I have located a group of injury claims which are have interrelated participants through link analysis. All of the individuals are associated with one and other through various unique identifiers contained in my data such as address, phone or vehicle. The link analysis chart is the first part of my presentation to the investigators, now I need to incorporate a geospatial visualization to my presentation to provide investigators with the locations used by the ring to execute the fraud.
















For this analysis I am going to use an analytical program called Centrifuge, which deploys a tool that allows for the integration of data with the Google Earth program, allowing for the mass visualization of geospatial data contained in my claim system.

While I know that all of the parties involved in this large group of injury claims are all interrelated, geospatial analysis is going to allow me to visualize their proximity to one and other. I can also incorporate the staged accident locations to provide a visual representation of the losses to the claimants addresses to further substantiate my theory of organized fraud activity.

I am going to download the claimants location information and loss locations from my claims database into an excel spreadsheet for upload into Centrifuge. Next, through the programs user interface I am going to select the columns that contain the claimants addresses so that Centrifuge can send the data to Google Earth for the visual representation.

From the example above you can see I selected the claimants street address, city and state. Centrifuge also allows me to bring in descriptive data for Google Earth, so I am selecting the claim number, claimants name and loss date to assist me with investigative planning.













After I have selected all of the data I wish Centrifuge to import into Google Earth, I click the show map button on the Centrifuge interface and the data is imported into the Google Earth program.















From this first perspective at the city level in Google Earth, you can see that the claimants are clustered together within the Orleans Parish. There are a couple of outliers which I am going to want to investigate further to ensure their connection with the ring, however my geospatial analysis is matching up with my link analysis, showing a cluster interrelated activity.

Another very useful feature of the Centrifuge tools import is located on the left menu of Google Earth. When I set up my import from the Centrifuge console, I also imported in the claimants name and date of loss as descriptors. On the left menu, you see that the claimants names are listed in Google Earth, if I or the investigator want to quickly locate an individual suspect in my staged accident ring, I just need to find the claimants name in the left hand menu and click on the hyperlink, Google Earth will then navigate to that location.













Lets drill down a bit into my geospatial analysis, from the visualization below you can see that we have four separate claimants all living less then 1 mile apart from each other. When I hover over one of the location, we can see the total number of claims where this address was used as the claimants residence.


















Staged accident rings typically need a body shop, legal provider and medical provider to maximize their return on investment so I need to incorporate the providers locations to my analysis as well.


















In addition to providing a link analysis chart to the investigators showing the interrelation between the claimants and the claims, I can now add a geospatial analysis to my presentation showing that all of the interrelated claimants live within one mile of each other and all in proximity to the legal provider, medical provider and body shop which are servicing the rings claims.

As an added benefit to using the Centrifuge tool and Google Earth, I can email or save the Google Earth file and provide it my investigators along with the location menu so they can quickly identify the location of the claimants they are going to investigate.

By also incorporating the loss locations into my geospatial analysis, I can further assist the investigators and law enforcement when developing an investigation and surveillance plan to catch the ring staging accidents.

Medical Provider Fraud

In the last example we incorporated medical providers into a geospatial analysis of staged accident activity to show the close proximity of the medical provider to the claimant. In medical provider fraud through, sometimes the opposite is a strong indicator of fraud.

We can leverage geospatial analysis to show inconsistent patterns in distance between claimants and a specific medical provider which could be an indicator of medical provider fraud.

By examining normal medical provider to claimant location records during any given period, the majority of individuals who treat with a medical provider in an accident claim, treat within five miles of the residence. The average distance changes based on location and population density so it is important to establish a clear average before performing geospatial analysis to locate potential medical fraud, but for the most part it is common sense that if you are injured in an accident, you are going to seek out a provider that is in an immediate proximity to where you live.

By incorporating geospatial analysis of a particular area, I can visualize clusters of injury claims where there is an abnormal distance between a group of treating claimants and the medical provider which would be a strong indication of organized fraud activity or medical billing fraud such as billing for services not rendered.

To look perform geospatial analysis of the distance between provider and claimants on a large scale will require an amp'ed up mapping program such as ArcGis, however on a smaller scale we can use Google Earth for specific proactive investigations.

In this example I located a medical provider through data mining who is treating a large number of vehicle injury claimants with an increased velocity of certain CPT codes which indicate fraud. I am going to utilize Centrifuge to import the claimants locations into Google Earth to show the relative distance between the provider and the claimants to substantiate my suspicion of fraud.














From this visualization, I can see that the claimants, for the most part all live in Jefferson and Orleans Parish, in and around metro New Orleans. From my medical billing database, I am going to use Centrifuge to import in the billing medical providers treating locations from their HICFA forms.














From my geospatial analysis you can see that the medical provider is located many miles away from the largest cluster of claimant addresses. To determine if there is a practical reason why so many people would travel such a distance for treatment I am going to use Centrifuge to import in all Chiropractors which have billed my company in the past six months in this region.

As you can see from the illustration, there are numerous providers in close proximity of the claimants which are not being used by this cluster of claims (represented by the red cross). Through the use of geospatial analysis, I have established that average distance between the claimants and ABC Chiropractic is over 10 miles and that there are 12 other medical providers of the same discipline in much closer proximity.

While this does not confirm fraud, it is the first step in the proactive identification of potential medical provider fraud. The geospatial analysis will be coupled with a velocity analysis of CPT codes being billed, a time line of patient treatment and a link analysis between the actual claimants to determine the probability of fraud for referral to investigation.

The analysis of the CPT codes being billed can also be accomplished in Centrifuge using the chart function after uploading my billing data into the program as illustrated below:





















I can utilize the timeline function in the Centrifuge program to visualize patterns to the treatment dates of the claimants. From this visualization I can establish if the patients are all treating on the same days, pattern of days, pattern of hours and duration.














By leveraging geospatial analysis I have proactively identified a medical provider who is treating a bulk of patients outside of an average proximity of their residence. Then by leveraging the visual analytical tools in Centrifuge I have visualized the medical billing to produce a presentation of potential medical fraud for my investigation team.

This can be accomplished on a much larger scale by utilizing specialized mapping software such as ArcGis to view in bulk the geospatial relationships of thousands of medical providers to their patients.

Catastrophe Claim Analysis

Geospatial analysis is critical in the proactive identification of catastrophe claims. Catastrophe claims pose a significant challenge to insurance companies as they produce large volumes of claims which under most statutory regulations, have to be handled in a short period of time.

The penetration of SIU into potentially fraudulent catastrophe claims has to occur almost simultaneously to the time the claims are being adjusted to prevent payment and exposure to the company.

By utilizing geospatial analysis, we can overlay the claim location with geocentric weather or disaster data provided by the national weather service or noaa to pinpoint claims with a loss location outside the path of the specified catastrophe.

Lets take an example of a tornado hitting southern Louisiana. I am going to leverage geospatial analysis to import the location of catastrophe claims my company has received with the corresponding storm track information to identify those claims which were outside of the reported storm track.
















In my first import, I can visualize the loss location of all CAT claims associated with this specific event. My next step is to import the storm track data from the national weather service to determine which claims were outside the path of this specific storm.

From the illustration below, you can see that the majority of the CAT claims fell within the storm track and damage geo data from the national weather service. There are several claims with a reported loss location outside this area that I will bring in for investigation.

















These are a few examples of effectively incorporating geospatial analysis into insurance fraud investigation and identification. Geospatial analysis, coupled with link analysis, provides a complete analytical representation of potential fraudulent activity that can greatly assist in the planning and execution of insurance fraud investigations.

I would like to thank Centrifuge for providing access to their analytical software for this article. For more information on their products please visit Centrifuge Interactive Analytics.