Adding more powerful map visuals to Power BI with Azure Maps, including custom tiles and real-time traffic, is just the first step to working with geodata for data visualisation.
Over the five years that Microsoft’s Power BI visualisation service has been available, it’s added a wide range of visualisations but the built-in mapping visualisation has remained fairly basic. The new Azure Maps visual for Power BI (currently in preview: enable it in the Power BI desktop app under File > Options and settings > Preview features) adds richer maps visualisations to the built-in list; but it also marks the beginning of much stronger geospatial analysis capabilities in Power BI generally.
“Our goal is to increase the geospatial analysis capabilities of Power BI out of the box. That’s one of the highest requested features for Power BI and this visual is our starting point,” explains Azure Maps principal technical program manager Ricky Brundritt. “We’re going to keep building on it to bring more and more data visualisation capabilities and analysis capabilities.”
“We also then have plans to go beyond the visual and start looking at other things that we can start tying into the Analysis Services that Power BI is offering geospatial analysis services through there. And then, even going further and looking at the Common Data Model in the Power Platform (CDM). Today, if you want a location, you might have a latitude longitude field or a place name, but there is no real location object inside the CDM. So that’s something we’re working with them to enable in the future, so location is a first party citizen from start to finish.”
Map data as bubbles and bars
Pick the new Azure Maps visual from the Power BI toolbox for a data set that includes geodata and you can see that data on a map as bubbles that show the scale of metrics like sales volume for that location. Initially, you can add latitude and longitude columns to your data set; when it’s closer to GA, the Azure Maps visual will also support address fields for geocoding – including the option to limit that to a single country for greater accuracy.
Like the built-in map visual, the bubbles use linear scaling by default but you can switch to logarithmic or cubic scaling (a popular request from Power BI customers) to make the differences in scale more obvious.
You can also pass in a value to set the colour of bubbles dynamically, so they can show category as well as scale. If you’re putting bubbles on a map to show accident hot spots, you can use the colour to show average speed and size to show how many people were involved; or you could show the category of the top-selling product in each area as well as the price range. In the future, the map visual will support Power BI’s conditional colour formatting to control colours, Brundritt says.
Documentation: Getting started with the Azure Maps Power BI visual
Documentation: Use the Azure Maps visual
You can also manage the zoom level and scale, to control when the bubbles are displayed as you zoom in and out, and whether they stay large enough to see when you zoom further out.
To make sure the data stands out on the map, pick the high-contrast option for either the bubble outline or fill and Power BI picks a colour variant that will be clearly visible, even on a dark map background. If you’re familiar with the Azure Maps Web SDK, you’ll recognise options like blur. “If you have a very large set of data points on the map you can use this as a way to create something like a heat map,” Brundritt explained.
If you want to tilt the map, you can control the pitch of bubbles and whether they lay flat on the map; that’s more important when you switch to 3D charts that show data as cylinders (the default) or bars. You can even switch between bubbles and bars automatically as you zoom in and out; if you’re looking at an entire country, bubbles give you an overview and then as you zoom in to city or state level, switching to a raised bar chart gives you more detail to compare at the local level.
If analysts are going to want to focus in on one bar of data at a time, turn on ‘unselected transparency’ and when they pick one shape, the others can fade out slightly (or completely).
Handling more data
Although you can work with very large geodata sets, Power BI loads the first 30,000 rows of data (although performance is now much improved working with large amounts of data). If your data includes a wide range of values, those first rows might not contain true minimum and maximum data values, so to avoid having to rescale the bubbles or bars for new data you can set the minimum and maximum values directly, and choose whether to show negative and zero values on the map.
Although Power BI can only connect a visual to a single data set, for maps you will often want to show two sets of data to show how they correlate: colour-coding counties or census districts by population or income values can help explain figures for hospital admissions, educational achievements, house prices, accident levels and other demographics.
In the long term, the Azure Maps and Power BI teams are working to find a way to connect a visual to two data sets, but the reference layer works around the limitation by letting you upload a GeoJSON file with location data and overlay it on the map using style properties to customise the look.
The only common GeoJSON styling option you can’t use is loading arbitrary image files (to avoid bringing external files into Power BI for security and compliance reasons). You can also overlay custom tile layers. These can come from your own tile service or use one of the additional Azure Maps tile services like weather radar.
Geodata as insight
In the longer term, location-related data like weather might be available beyond maps, along with other historical data like census data and historical traffic, so that you can use in analysis whether you need to see it on a map or not. If you’re making sales predictions or working out insurance risk, weather can be a significant factor in what will sell where or what accidents might happen, but seeing that weather on a map isn’t necessarily helpful as part of doing those predictive analytics.
Making that information available in a self-service business intelligence and data visualisation service like Power BI makes it accessible to far more people than if it’s only in developer tools. Adding the Azure Maps visual to Power BI starts by updating the built-in map visualisation but in the long term, it makes geodata another source of insight.
For now, it’s still a preview though. There’s more work to be done on performance and to allow loading even larger data sets, as well as more features under consideration for the map visual – like the possibility of handing well-known text as well as addresses for geocoding, or supporting lines and polygons so you can draw custom boundaries. You can’t currently publish Azure Maps to the web with Power BI or embed it, but that will change once Microsoft has confirmed costs with the third-party data providers used in Azure Maps.