AI and Location Analytics: How is AI enhancing indoor location data analysis?

pexels-googledeepmind-17483868For many of the companies that invest in an indoor positioning system or indoor maps, the end user experience is the key driving force behind that investment - companies want to give the staff, customers, visitors or guests a rich, rewarding experience by helping them get where they need to go, find their friends or co-workers, alert them to nearby deals, or literally one of thousands of other use cases unlocked through the power of indoor location.

An additional bonus on top of the fantastic experiences enjoyed by end users of indoor location systems is the potential for businesses to gain access to hitherto hidden insights into user behavior. Analytics tied to indoor location systems can tell us everything from the most popular locations within a retail store, to common congestion spots within an airport, to warehouse areas that are underutilized.

Now, with the monumental advances being made in AI, businesses across numerous industries and verticals are not only being provided with huge amounts of data from their indoor location providers, but have a way of synthesizing it into invaluable insights quickly and with minimal additional analysis. In this post, we’ll explore ways in which the AI boom is impacting the indoor location analytics market and some of the exciting use cases that have already emerged.

What are some popular indoor location data points? 

Before we dive in to look at AI’s impact, it’s important to appreciate some of the most popular data points commonly tied to indoor location.

General visit data 

While not as flashy as some of our later data points, it’s impossible to overstate just how critical basic visit data can be to a business. Analyzing data such as number of app users or visitors over a set period of time can tell a business everything from when they ought to be putting a greater number of staff on the shop floor to deal with increased demand to the efficacy of a recent marketing or brand awareness advertising campaign

Dwell time

Much non-indoor location based data is binary, and leaves little room for nuance. Two airport visitors could travel via the same terminal and catch the same flight; however, one could have gone straight from bag check, through security, and to the gate, while the other could have spent hours browsing duty free stores and eating at restaurants before arriving at the gate and boarding the plane. This is valuable information for airports and can help inform their strategies to increase non-aeronautical revenue, based on which types of passengers are dwelling in different areas of the airport.

pexels-wdnet-106344Search data

Using purchase or interaction data to analyze a particular product or service’s success can be effective - but it can also be deceptive. Take for example a store which is supposed to have a product, but that has run out of stock. If sales data was then analyzed by someone without the full picture (i.e. the knowledge that the product was unavailable), the assumption could be made that there was no demand for the product. However, by using in-app search data, the ability to see what shoppers were looking for regardless of whether they were then able to buy it is unlocked, and with it a vital insight into potential opportunities within the supply chain. The data could even be used for products that the store has never stocked, to reveal products or verticals they should be investing in.

POI interactions

Somewhat similarly to search data, enabling businesses to see which points of interest (POIs) in their location are proving to be most (or least) popular can provide telling insights to how they could optimize locations, and can shed further light on insights gleaned from basic visit data. To use another retail example, a clothing store might have a women's, a men's, and a children's department. Using only visit data would provide limited insights. However, using POI data, you could tell which of these three departments were drawing the visitors into the store in the first place, and which were having a negligible impact on the store’s general appeal.

Popular destinations and journeys

This is truly the data where an indoor location comes into its own and can deliver insights that no other system can. Analyzing basic data enables airports to see flight occupancy data, retailers to see sales figures per product, or workplaces to count employees in or out. 

However, datapoints such as these only speak to one final result, and leave little room for interpreting how that conclusion was achieved. Did each flier walk straight through to the gate, or did some give the airport more in non-aeronautical revenue than they did with their flight itself by buying gifts in multiple stores? Did the retail shopper know exactly what they were looking for, or did they spend time browsing and comparing before purchasing? Did an employee lose valuable time in the working day hunting for a spare desk or meeting room?

None of these insights are possible without being able to analyze a user’s physical movements throughout the real-world environment of your building - and these are only scratching the surface of what’s possible to see when it comes to destination and journey data.

How is AI impacting indoor location analytics?

As you will have seen from the above section, there is no one single panacea when it comes to using location data. General visit data is useful in some contexts, but doesn’t contain much in the way of granular detail. POI or search information can be extremely valuable, but can be influenced or distorted, a trend which may only be uncovered by combining this data with destination and journey data (e.g. fewer users will need to search for a POI if it’s immediately obvious to the naked eye when one first arrives in a location, compared to a POI that’s harder to find). In short, the art of indoor location data analytics only begins with the data itself - the truly valuable insights are gained by the ability to synthesize data points of different types, understand how they correspond to and influence one another, and develop insights and judgements based on this.

pexels-tara-winstead-8386440Lets revisit our trusty retail store example, and take a journey through the analytical process that one looking to analyze the data might undertake:

  • One day, there’s an unusual high number of visits to a store (indicated by basic usage data, the number of in-app searches and wayfinding journeys, etc.)

  • The store has recently run a promotional campaign around a particular type of products - using search, dwell, and wayfinding data, we can check whether more foot traffic was concentrated towards the promotional products
    • If it was, it’s a valuable insight into the efficacy of the promotional campaign

    • If the traffic doesn’t seem to be particularly focused on the promoted products, then it can be inferred that the promotion wasn’t the reason for the increased traffic, and that there may be another cause. If so, the data can be further scrutinized to see if there was one clear reason for the influx of visitors, or if it just happened to be an unusually busy day
  • In retail, it’s not just about one-off purchases - repeat business and basket size are critical. Using app data, the retailer would be able to tell how many of the visitors that day had visited before, how many were first time visitors who eventually came back, and how many of the visitors made a beeline for the promoted product and then checked out, versus how many spent additional time in the store

This one example shows just how many considerations and factors are at play in one single piece of analysis. There’s also considerable room for human error, either via misinterpreting or misreporting a piece of data, or by simply not considering all the possibilities locked within the data; using our example above, it would be very easy to make the connection between a higher number of visitors meaning the promotional campaign was an unmitigated success and leaving it at that.

This is where AI promises to be such a game-changing innovation. Not only do we now have data available that previously didn’t exist, but we now have a tool in AI that is capable of ingesting multiple pieces of data across a significant time period, detecting fluctuations or notable changes in the data trends, and combining data points together with a speed and precision rarely possible via human analysts. 

pexels-pixabay-265087AI indoor analytics use cases

With AI’s aforementioned ability to monitor huge quantities of data as it’s provided, it’s enabling exciting new use cases for data across a range of industries:


Airports are some of the most dynamic indoor environments there are, with hundreds or even thousands of people, each with a set destination and time to get there, moving through tightly controlled areas each hour. Understanding bottlenecks and seeing where visitors may be having issues moving through the airport in real-time is critical to the effective running of an airport and a positive visitor experience for fliers. Using AI with occupancy data, airports can not only spot bottlenecks or crowded areas as they happen, but can use predictive models to be alerted to them before they even happen. AI can detect the patterns that tend to occur in the build-up to an overcrowding event - for example at a particular baggage carousel or security desk - and alert operatives that more staff are needed or another lane needs opening before a blockage has even occurred.


With events such as trade shows generally taking place over a small number of days, time is of the essence when it comes to identifying issues as they occur. This is something an AI-enabled indoor location analytics system would excel at. By monitoring real-time data, an AI can warn event staff when a particular area is becoming overly busy, prompting them to help disperse crowds, open alternative pathways, or advise newcomers to the area of different routes to avoid the congestion.


Healthcare facilities have already been revolutionized by indoor positioning systems, which have helped guide guests through complex facilities and reduce appointment no-shows due to visitors getting lost. Indoor positioning is also critical for asset tracking systems, which can help hospitals keep tabs on expensive pieces of portable medical equipment. Using AI, a model can be built of how effectively a hospital is utilizing these key pieces of equipment, and be used to inform everything from the best places for these devices to be stored so as to ensure maximum availability for the widest number of staff, to helping allocate more budget to the most over-subscribed pieces of equipment.


Hospitality is an area where AI-powered location analytics is particularly useful, as many of the actions that visitors may take will leave no tangible trace like a purchase history or defined goal (such as boarding a plane), meaning that actually measuring ‘success’ and finding trends that link high value visitors is harder. Take for example a cruise ship. Tying user data and location data together, AI-powered analytics could take several cruises’ worth of data to detect which user demographics are making the best use of the on-board entertainment and facilities, and which demographics are using fewer of the ship’s facilities. In order to reduce customer churn and encourage repeat business, the cruise line could then target the demographic that traditionally makes less of the on-board facilities with in-app notifications aimed at encouraging them to try activities and thus improve their experience, and the chances of them returning in future.


As has already been extensively discussed, retail is an industry in which much data was already available, but which often lacked context only available via an indoor location system. Two customers can appear identical if they buy the same items; however, understanding the difference between one who walked directly to that item, then straight to the checkout, and one who spent time browsing and comparing products, for example, is an absolutely critical retail insight that could be utilized in a vast number of ways. AI can help spot the trends and tendencies, helping retailers identify which visitors were a one-off, and which could potentially become high value repeat customers, among literally thousands of other possible insights.


Workplaces were some of the first building types to truly embrace the transformative potential of indoor location, and they remain at the vanguard of some of the most cutting-edge insights that can be gleaned by combining indoor positioning systems with AI-enabled location analytics. AI means that businesses can quickly garner insights in terms of building occupancy, which can help in discussions about whether to up- or downsize office real estate, and areas of the office that are frequently under- or over-utilized, right through to calculating how much energy (and money) are wasted by HVAC and lighting systems running in rooms with nobody in them.

Interested in learning more?

Pointr's AI-enabled mapping engine combines with our bespoke analytics platform to deliver world-class location experiences for visitors and incisive business insights for clients. Let us show you how we can help your business.

Contact us