Abstract

People interact with geospatial data every day. Some use it to make critical business decisions, while others are simply trying to decide where to eat out for dinner one night. In any case, it’s crucial for this data to be correct in order for people to confidently make choices involving what places are in the world, and where.

But getting geospatial data right isn’t as easy as it may first sound. There are so many different points of interest (POIs) worldwide to track, and many of them can change rather quickly. New places can open, failing places can close, and even businesses going steady may change their operating hours, contact information, or branding.

In this webinar, SafeGraph geographer Briana Brown outlines examples of use cases in which having accurate geospatial data is paramount. She also details some conventions geospatial data providers can follow to help them maintain their data’s precision, despite frequent changes to different POIs around the world. Here’s a summary of what’s in store:

• Why accurate geospatial data is so important
• 5 key takeaways from this webinar about accurate geospatial analysis
• Point POIs vs POIs with Polygons
• How SafeGraph data enables accurate geospatial analysis in a constantly changing world

We’ll start with a quick discussion of some ways geospatial data is used, and why it’s critical to get the data right in these situations.

Why accurate geospatial data is so important

One reason why geospatial data needs to be right is because of how often it’s used in modern daily life. Think about the kinds of information you’d look for when opening a mapping application on your phone. What restaurants are nearby, and what kinds of food do they serve? Where can you find an ATM if you get low on cash? Where is that shop you want to visit, and when is it open for business? Is that shop even still operating, or has it closed permanently? These are all questions that require accurate geospatial data to answer.

That’s to say nothing of the benefits of using geospatial data in analytics. When curated properly, geospatial data allows for examining the relationships between places, industries, and brands from multiple angles. For example, you can survey a brand’s overall footprint in an area while taking into account the contexts of that brand’s individual stores. Or you can scan an area to ascertain which kinds of businesses and services are prominent there, and which might be needed but are missing. Or you may notice certain types of businesses are rapidly closing in one area but opening in great numbers in another area, perhaps pointing to a broader socio-economic change.

In any case, it’s important for geospatial data companies to not compromise on either the accuracy or currency of their data. Not only are there a lot of POIs out there to get the right information about, but they are also constantly changing in various ways. Some are newly opening, others are closing permanently, still others are being rebranded, and so on.

Out-of-date or otherwise inaccurate geospatial data in consumer-facing applications can lead to bad user experiences, higher customer churn rates, and a negative reputation for the organization. Similarly, there can be significant consequences for misinformed corporate decisions based on flawed geospatial data: wasted advertisement spending, mistargeted marketing strategies, missed investment opportunities, and inefficient product distribution networks.

5 key takeaways from this webinar about accurate geospatial analysis

We’ve given some brief examples of why geospatial data is important to get right, as well as a couple of reasons why that’s challenging. So the following points will explain some of the key principles SafeGraph follows in order to offer accurate geospatial data as a service.

1. We’re expanding our Places dataset because data scientists need a single, comprehensive global POI dataset 

Time in Video: 3:15

As of October 2021, we’re now documenting data on POIs around the world – instead of in a few specific countries – in our Places dataset. That includes POIs for both large multinational corporate chains and smaller local brands. Our goal in doing this is to improve our product to meet the demands of today’s data scientists.

There are at least two reasons why we feel this expansion is necessary. First, as the world becomes more globalized and interdependent, analysts are often required to compare data on multiple regions of the world at once. Being able to do so with a single dataset makes it much more efficient and cost-effective.

Second, having to analyze POI data from multiple separate geospatial data sources can create certain problems for data scientists. If these sources have different formatting conventions, or even different data about the same POIs, data scientists have to do a lot of clean-up work to make the data usable. This can involve a series of judgment calls regarding both standards for how to represent information about POIs, as well as what information about POIs is relevant or accurate. Arbitrariness with the data can result in analyses that aren’t very trustworthy.

2. Openings and closings are important POI metadata, but difficult to track

Time in Video: 7:03

Maintaining precise data on when POIs open or close is crucial for fields like urban planning, real estate development, and financial analysis. People in these industries need this data to monitor how brands, industries, or communities grow or shrink over time. This allows them to identify areas of opportunity for investment, or to notice when certain communities are being underserved. The problem, however, is businesses can open and close rather abruptly. For example, under normal circumstances, about ⅕ of all new businesses close within a year of opening.

We at SafeGraph have developed techniques to flag POIs that repeatedly disappear or newly appear in our data pipeline as closed or open, respectively. Out of transparency, we also differentiate between when we determine a POI closed and when we started reporting it as being closed, in case this is relevant information for our clients’ analyses or reporting. We update our Places dataset monthly to ensure this information is as precise as it can be.

See our blog or our Places dataset documentation for more information.

3. Apartment-like buildings are POIs because of their business relevance

Time in Video: 8:44

For privacy reasons, SafeGraph does not count single-family homes or individual apartment units as POIs. However, we consider buildings housing multiple families – such as townhouses, condominiums, and apartment complexes – to be POIs. One reason why is these places are technically their own type of business: they rent or lease living space to people.

Another major reason is advertising opportunities – such as posters or interactive TVs in lobbies and elevators – are increasing at apartment-class POIs. This is making data on apartment-class POIs very attractive to some marketers, as they look to take advantage of new advertising spaces that bridge the gap between in-home and out-of-home marketing. 

However, this data must be kept up-to-date to be useful. The demographics of people who live in apartment-class POIs are changing, as many people are leaving downtown city cores to live in suburbs. The types of other places surrounding apartment-class POIs are changing as well. So if this data is stale, marketers can end up wasting ad spend targeting the wrong audiences at the wrong places.

4. Data on POIs without polygons can still be useful for marketing and urban planning

Time in Video: 10:11

Some POIs do not have traditional geometric building shapes. However, they are still relevant because people interact with them in their day-to-day lives. We refer to places such as electric vehicle charging stations, vending machines, ATMs, and transit stops as “point POIs”.

Data on these types of POIs can be useful to marketers, as these places often present similar opportunities for out-of-home marketing as apartment-class POIs. Point POI data is also valuable for urban planners, letting them examine whether public transportation and other amenities are available in the right places and quantities to meet community demand. It also allows them to assess if the current transportation infrastructure provides accessibility to critical services (e.g. grocery stores, hospitals) for as many people as possible.

5. Industrial POIs are becoming more important for studying supply chains

Time in Video: 11:44

There is increasing demand for geospatial information about industrial, non-consumer POIs. These include warehouses, self-storage units, manufacturing plants, and data centers. Organizations need this data to conduct economic research on global supply chains, and/or plan their own production and distribution processes. This data helps to answer questions regarding raw material availability, where a product is in the distribution process, logistics, warehousing capacity, and regional consumer demand.

Events like the COVID-19 pandemic have made global supply chains extremely volatile, so having up-to-date information on industrial POIs is more critical than ever. Unfortunately, many POI datasets are not updated frequently enough – often quarterly or annually at most. So analyses that use them risk modeling locations that no longer exist, or failing to model newly-opened locations.

Point POIs vs. POIs with Polygons

For most of the POIs we catalog data on, we use polygons to delineate their physical shapes. You can find this information in our Geometry dataset. However, we also catalog data on POIs whose shapes are difficult or impossible to describe accurately on a map with polygons. We refer to these as “point POIs”. They include things like ATMs, transit stops, vending machines, and electric vehicle charging stations.

We compile data on these sorts of places because they are still significant locations people interact with on a daily basis. Many are extensions of certain businesses or brands, or can present out-of-home advertising opportunities (such as posters in bus shelters). And they can change just as frequently as traditional buildings do, if not more so.

This sort of information is highly sought after by urban planners. It helps them examine if communities have sufficient access to public transportation and other services. For instance, it can be used to check whether businesses have certain amenities available, in terms of both accessibility and potential incentives (tax breaks, for example) for including them. It can also help local governments see if there are critical services that lack sufficient accessibility via public transportation. Finally, it can shed light on whether the presence of certain types of amenities aligns with community demand for them.

How SafeGraph data enables accurate geospatial analysis in a constantly changing world

SafeGraph aims to be one of the leading sources of geospatial data on the planet. Our flagship dataset is Places, in which we catalog information about millions of points of interest across the globe. We also update this dataset monthly so organizations can work with reliable data on which places open, close, change, or currently exist.

We build our other datasets based on the information in Places. Our Geometry dataset, for example, outlines the physical shapes and sizes of POIs with polygons. It also contains spatial hierarchy metadata that denotes the relationship between two or more places whose polygons overlap (such as a tourism center within a park or a store enclosed within a mall).