Use Case: How our Polygon Data provide valuable insight to an insurance company

November 10, 2022
Written by Barnana Sarkar

Polygons around a POI mark out the exact location. Source: Echo internal stock.

Polygons or virtual boundaries are drawn around a particular Point-of-Interest (POI) and act as the perfect taper to filter out information – such as location, building type, shape, or measuring the proximity of two POIs – a business needs to succeed. Polygons not only represent the physical features of a piece of land or water body, such as the height of the landscape or the depth of the lake but also act as effective risk estimation models for several industries like the insurance industry.

At Echo Analytics, we always believed in polygons’ potential to determine an area’s facets. Our client, an insurance company, was interested in understanding an industrial complex area to estimate their investment risks. We recognized this opportunity to enrich their already existing POI datasets by combining them with polygons to give them an understanding of the location characteristics, such as the number of residential buildings, the mobility trend of the area, and if there are any upcoming projects in the area.

The Challenge

Our client – an insurance company – insured an industrial park for a minimum value of 100 million eurors but they soon realised that they were facing a loss on revenue. They struggled to estimate the exact reason behind it and to determine that cause. They approached us with their already existing POI dataset of the industrial park. Our first challenge was to enrich that dataset with our polygon dataset.

A POI dataset that is enriched by polygon data provides a deeper and wider understanding of a property’s location since a POI alone might not be entirely accurate. There might be situations where the data shows that the POI is outside the risk zone. Our client, despite having a surface-level dataset of the industrial complex, needed an in-depth understanding of the features of that location such as whether the complex was located in a high-risk or low-risk zone. A location is said to be a high-risk zone when it has more chances of getting exposed to any kind of danger or – in case of a residential area – the residents are at a higher probability of facing the consequences of a calamity. Similarly, a low risk zone has lower chances of being impacted by any natural disaster or calamity. Insurance companies must have an accurate and precise understanding of a particular location for proper risk estimation, and an enriched dataset with polygons can help them do so.

However, even for expert data analysts – such as Echo Analytics – sometimes distinguishing the high-risk zone from the low-risk zone might be difficult because they often overlap. Let’s take the example of the image below. The map below shows the flood-prone coastline along the Gulf of Mexico. Each zone is distinguished by a prime color, such as the areas that are at maximum risk of flooding are marked in red and the comparatively safer areas are marked in blue. However, when we look closer (such as the section highlighted in the circle) we notice that the colors begin to overlap. This is a common case for any polygons which are usually drawn over an extensive area such as the coastline that we see on the map. Although they might be precise, they can make it difficult to mark a location accurately. This is a challenge that we faced head-on within the case of our client.

In most maps, polygons might overlap therefore marring the purpose of distinguishing one area from the other. At Echo Analytics we ensure that this is avoided.

Source: National Hurricane Center and Central Pacific Hurricane Center.
Downloaded on November 3 2022.

Our objective with our client was to mark out the areas in the industrial complex which were at high risk of flood and fire. While creating polygons around the POIs that we attained from their datasets, we noticed that there were several low-risk and high-risk areas which overlapped. This was indeed a challenge but we tackled it head-on. We split out the addresses that were already available on our clients POI’s datasets and began to mark out each address with polygons. Once we did so, we were able to inspect within 1 km of each listed address and marked out the high-risk zones from the low-risk ones. This provided our client, the insurance company, with the scope to automate their policy setting that could cater to their customers as well as ensure increase in their revenue.

We know that the more people crowd in an area, the higher are the risks for human-inflicted damage – like a stampeded at a festival. We wanted to help our client further estimate the visit attribution through which they could take necessary precautions such as requiring the building owner to draft an evacuation plan in the case of a fire or have lifevests for people in the case of a flood. Once these visit attributions were integrated into the polygons, our client was able to get a detailed informed picture of the industrial park. This resulted in obtaining 360 view into visits pattern and increasing safety in the industrial park.

Our takeaway

It was interesting to note how the right information about a physical space not only determines a company’s revenue but also the overall welfare of the people that the company serves. In the case, of the insurance company – our client – it became much easier for them to predict a mishap once they have the polygon data with them. When they could predict the accidents, they could make better investment decisions as well as take measures to ensure the protection of the residents around the industrial park.