GB : What is 'Location Intelligence'?

Published: Monday September 04, 2017

What is Location Intelligence? Our Data Scientist, Dr Ian Hopkinson, explains more.

Customers hate filling in forms. The more questions they are asked, the more likely they are to drop out of a process. This leads to loss of business. But what if you didn't need to ask your customers questions because you already knew the answers?

This is where Location Intelligence comes in. It grew from the realisation that datasets recently made publicly available, combined with the appropriate commercial datasets, provided much of the information required for customers to apply for buildings insurance. As an industry with revenues of £3.6billion per year (figure based on ABI key facts), location intelligence is key to ensure smooth customer experiences.

Buildings insurers typically ask for information such as the property type (detached, semi-detached, terraced or flat), number of bedrooms, number of bathrooms, number of floors, and property age. This data enables them to estimate the risk a property presents and therefore calculate premiums accordingly. Location Intelligence can supply this data across all properties in England, Wales and Scotland.

We're also able to estimate a property footprint and the area of the building plot it sits on, as well as its height as determined from the Environment Agency LIDAR data. We can provide an indication as to whether the building is listed which has an enormous impact on rebuild costs, an estimate of current value and an indication of the burglary rate in the neighbourhood.

Many of these attributes have long been used in the buildings insurance industry but items such as the building footprint and height are new. Location Intelligence allows insurers to conduct 'what-if' analysis on their historic data to establish the business value of these new fields.

But Location Intelligence isn't just of value to the buildings insurance industry.

Where else can we find Location Intelligence in practice?

Companies making deliveries or doing installation work can also benefit from this data, since it can provide a measure of how difficult a delivery might be or whether any special equipment might be required. We're also providing information on red routes, the congestion zone and the distance of a property from the road.

Companies engaged in trace and investigation activities can check whether what they know about a subject (a debtor or hirer) of an enquiry is consistent with where they are purported to live.

Companies planning new facilities or infrastructure can use Location Intelligence to build a detailed picture of a neighbourhood or wider area.

Each of these applications provide direct business value, in improved click-through rates, in reduced failures to deliver or install, in enhanced investigation and in analysis. Whilst much of the data is available openly, building the entire dataset and maintaining it is a significant effort - can you afford to have one of your team dedicating themselves to this, rather than carrying out the core analysis functions of your business?

So how is it done?

Building Location Intelligence is not simply a case of shovelling a large collection of datasets into one large file. With colleagues in the buildings insurance industry we collected a database of property details for 70,000 properties, a so called 'ground-truth'. We then evaluate each dataset against the 'ground-truth', standardising data where required. For each field and each property we provide source and 'confidence' information, to indicate where each data item came from and how accurate we believe it is.

Data that comes to us typically describes properties by an address. We use our address matching expertise to standardise these addresses, enabling us to cross-reference data from multiple sources for a single property.

Much of the data we use in building Location Intelligence is supplied as flat comma separated value (CSV) but some of it comes in the form of ESRI Shapefiles - geometric data which needs special spatial querying technology to derive the attributes we supply in Location Intelligence such as the distance from road and the building footprint.

Where we do not have direct data for a property, we will provide an estimate using various modelling techniques including machine learning and naïve Bayesian methods but you'll always know, using the sourcing and confidence fields, where the data we present came from and how confident we are in the answers.

Location Intelligence can be provided via our product GBG Matchcode360 or as a flat file.

If you're interested in Location Intelligence and want to find out more, get in touch at sales.enquiries@gbgplc.com