How to Get the Most Out of Point of Interest Data for Location Analysis

What is Point of Interest (POI) data?

POI data represents physical locations of interesting or useful places that attract people's attention. For example, a big supermarket, a pharmacy, a tourist attraction, an office building, or even that tiny hot dog stand on a train station. For us, regular people, the most common use case for POI data is when we use GPS or online maps to get to a specific location because we don't always know the particular address or coordinates. Instead, we just want to get to the nearest McDonalds or gas station. Why should you, as a business, care about the Points of Interest? Because it can help you with decision-making, connecting with customers and expansion strategies as it reveals the vital location context of specific areas.

Example of Point of Interest data use case 

A good example is site selection for a new branch of your business in the center of a city. Let's say you sell clothes and toys for babies named "Babycot." Ideally, you use POI to map influential factors on a potential site, such as competition in the area, traffic infrastructure, and other places relevant to your business, to make sure you choose the best location possible. However, doing location analysis only with raw POI data won't give you the perspective you need. After all, each point of interest has a different significance and attracts different kinds of people and different amounts of people. So how do we deal with that and make the POI data most accurate?

Grouping POIs into categories

Suppose you want to get the best of the POI data. In that case, it's ideal to divide them into categories -  collections of similar points of interest that contain more details within a category type. For our Babycot, there can be many relevant groups. For example - shops with clothes. While parents are buying clothes for themselves, they may want to buy something for their baby too. Also, kindergarten! Parents are picking up their children from the kindergarten, and there is a shop with baby clothes and toys right next to it. So they could stop by, right? Then the categories you would use here may be "Shopping - Clothes" and "Education - Kindergarten."If you get a whole set of POIs for the entire country, you need to consider that not all of them are relevant to your case. Thanks to the division into categories, you can easily select only those that make sense and avoid unnecessary expenses and database load.

Weighting the POI for the most accurate analysis

How does the weighting of POI work? Each point of interest gets a value according to its significance to the surroundings. So for example, a small grocery store on the street corner doesn't have the same value as a huge shopping mall with many shops, and a tiny hot dog stand on the train station does not attract crowds as McDonald's on the main square does, therefore gets a lower value. The picture below shows all the Points of Interest in the center of Prague visualized on a dot map representing where the POIs are located with no consideration of their size and influence.

Visualization of raw POI data.


In the picture below, we see the visualization of weighted POIs. The color represents the category and the size of the dot corresponds to the assigned value of each point. This gives us more precise information about the location.

Visualization of POI categories with assigned values.

At CleverMaps, we have developed our own algorithm for weighting the POIs and we are offering it as a solution called the Retail Exposure Index. We can read the Retail Exposure Index as reversed mobility data. Mobility data reveal the actual movement of people on a map, but the Retail Exposure Index tells us WHAT attracts people to go to specific locations. By weighing the points and grouping them into an index, we can get an estimation of the location's traffic for a far more reasonable price than with mobility data.

Where to get the POI data?

One of the most common POI data sources is open-source databases that upload a huge amount of data every day. For example, OpenStreet Maps or Gisgraphy. Providers often offer raw POI data, however, some providers can offer a complex solution directly. For example, The DataAppeal Company offers a wide range of solutions compiled from POI data including Sentiment and Market Intelligence. At CleverMaps, we crafted the Retail Exposure Index, which is ideal for site selection, expansion strategies, branch network optimization, and more. In the video below you can see how it works in more detail.