RGI Group
RGI Group -- _inspired by

Real Life Case in the Big Data era

The following scenario is an example of how new products for Non-Motor Branches of Insurance that can be created by using predictive models and making use of big data.

The Data Scientist is configuring a new "Multi-Risk" product for farms, combining existing business rules, calculations and models with the new variables from the world of Big Data.

Nowadays it is increasingly easy to track down a mass of useful information that directly concerns us. For example, it is possible to obtain through the network environmental information that is useful in describing and representing atmospheric phenomena. The main sources of data and metadata are:

  • Forest fires: risk index in particular zones
  • Zones with risk of hail: level of occurrence in specific areas
  • Zones with risk of flooding: % occurrence by zone
  • Zones with hydrogeological disturbance: localities and levels
  • Farms with significant risk: number of farms at risk by area
  • Zones with higher lightning risk: % probability in the reference territory

Let's look at their practical application...

Stefano is at his desk in front of his PC. Let's see how to define a new personalised "Multi-Risk" product that will cover losses due to environmental disasters with the lowest risk for our Company.

He creates a new product structure by setting up the version and the operating validity of the product, configuring the structure, and then he confirms.

He accesses the Dictionaries and he selects the risk "Farm Building".

To define the rules on whether the cover can be taken on, the Data Scientist selects the relevant factors.

It is possible to use a new exceptional source of information, the Big Data, which is available in the dictionary section and you can select those ones of interest simply dragging them into the Factors section.

First of all the data scientist selects the factors relating to the "Forest Fire Zones" with the related metadata, then "Areas" with hydrogeological disturbance and their metadata and so on for the other factors.

Before writing the rule, the data scientist starts the analysis of the incidence of these risk indicators on the territory.

Using the big data, the system shows the 6 areas with the lowest risk for the factors just selected. In details the Puglia region: close to Lecce we have the highest % of risk related to these indicators.

Let us go into the detail of the macro areas and the related zones...

This level of detail is particularly interesting. In fact the risk of hydrogeological disturbance is low in this macro area but the zone East of Lecce is an exception, where the risk of hydrogeological disturbance is very high.

This information is particularly useful in defining the rules of the new product on whether to take on the risk, and operating more accurately.

The data scientist returns to the rule engine and then he creates the new Rule.

He sets up the version, validity and description, configuring the conditions of the new rule.

Based on the data obtained for the Zone East of Lecce with a risk of hydrogeological disturbance of less than 30%, the criteria are set at "YES".

Another condition, still for the Zone East of Lecce if the risk of hydrogeological disturbance is greater than or equal to 30% the rule is set at "YES" with a maximum sum insured of 20% of [Value of property].

And then the data scientist saves the rule.

From the summary screenshot you can check the rule before the final confirm. Then you can save the new product.

"A potential insured goes to the Lecce agency to find an insurance policy that will protect her farm from meteorological risks and at the same time meets her budgetary requirements".

In an agency in Lecce, the owner of a B&B is looking for a product to cover her property, but she has a limited budget.

An Agent of the Company is working on his virtual desktop when the insured arrives at the agency and asks an help for her B&B.

She has a farm with an attractive B&B, but at the moment the farm has no insurance covers and she would like to look at the possibility of purchasing one with a limited budget, around €1.000. Her farm is in the countryside, in the province of Lecce, near a wood which has recently been affected by some small fires.

So the agent searches the best solution for a budget of €1.000, preparing the new quote.

He enters the quote of €1.000 and he inserts a B&B photo to make a quick and accurate search by image. The system is searching for the location of the B&B by the photo.

Here we are. For meteorological events, the risk of flooding or earthquakes in that area is sufficiently low, but the house is particularly exposed at fires.

There is also an expensive cover for landslides because the zone where the B&B is located has a high risk of landslides...

The insured understands that the landslide cover is necessary and the agent prepares the documents!