The quality of Point of Interest (POI) data is an important attribute to assess as it should provide powerful information to drive great business strategies. But what defines a good quality POI dataset? There are three important pillars of what entails a good POI dataset. These include:
- Recency
- Accuracy
- Scale
Truth be told, an ideal POI dataset should constantly evolve as our physical world changes every day. In 2020, thousands of businesses shut down in multiple countries, which resulted in millions of out-of-date POIs worldwide. To push leading-edge decision-making with this kind of data, companies like yours need to focus on leveraging a top scale of points of interest, as well as ensuring the recency and precision of datasets.
To assess the quality of POI data, companies have to run through the following factors:
1. Updated Points of Interest
Premium POI data should not only keep track of closed locations but should also include new businesses, neighbourhood expansions, new governmental and administrative facilities, and enrich all of the corresponding attributes for old and new POIs.
2. Precision and accuracy of the location
Accurate POI data entails making sure that locations do not duplicate nor do they include incorrect or outdated information for each of their attributes. Using incorrect information can lead to revenue losses or poor decision-making.
3. Ability to cross-reference data
According to Google, over 90 businesses change every hour (whether that means that they changed their name, location, size, contact information etc. or even shut down). This makes it challenging for any POI data provider to keep up to date with all the changes on an hourly basis. Multi-sourced data can be a great solution to obtain the most accurate and up-to-date information. To do so, it is important that the data that they leverage can be cross-referenced with multiple sources to enrich your scale of POIs and their corresponding information. It is also important to consider if you can request customisable datasets to fit all your requirements; you wouldn’t want to waste your time analysing information in vain.
4. Understand the collection method and get the right scale
In 2020, the U.S. government registered over 31M businesses operating around the country. To ensure that you get the right points of interest you should evaluate the way a provider collects the data and whether it covers interesting areas for your business.
5. Testing data samples before engaging on bigger projects
- Assessing the data at a neighbourhood or city level to check for density
- Checking a couple of brands per category on a nation-level to see the coverage per brand/category. Does the data match the statistics that these stores have online?
