This demo illustrates the use of KeyLines to discover patterns of fraud in typical insurance industry data.

The data-set represents the relations between typical actors in the car insurance domain. For example, vehicles, drivers and addresses all associated with a claim file. The demo emulates the workflow that might be offered in a real application.

After Claim 1 or Claim 2 is loaded, the user is offered the ability to 'Find Matches' between the entities shown and other data (for example, previously processed claims). The new data is then shown alongside the current claim's data. The demo then offers a 'Combine Matches' button that uses KeyLines' Combo feature to clearly expose the common entities.

While there is a common factor in the matched Claim 1 data ('Valois House'), this looks innocent as the incidents are over three years apart, and the Claim 2019-11-24 could be approved.

Claim 2's data matches show suspicious patterns related to vehicles, people and addresses, and this claim should probably be referred for further investigation.

In both cases the red Claim icon features a context menu option to select a further action.

The data set is fictional, but contains patterns typical of real fraud. Any resemblance to persons alive or dead is purely coincidental!

Key functions used:

• chart.layout
• chart.setProperties
• chart.show
• chart.combo().combine
• chart.combo().uncombine