With the September release of Microsoft Power BI Desktop, the recently announced ArcGIS Maps for Power BI are available in preview. After having an opportunity to use the Esri ArcGIS map for the first time in PBI Desktop, it is an impressive step up from what Power BI had before.
How impressive? Here’s how maps changed overnight in Power BI:
- You can now make beautiful and functional maps in Power BI. It is on par with standard fare from other BI tools like Tableau as well as a lot of basic customized maps you could make with open source tools.
- Alternative base layers are available. The default is now light gray. Someday I might clamor for alternative basemaps from Esri, but for now, dark and light options are a large step up for data visualization in Power BI. I can see my data without the Bing/HERE streets and colors getting in the way.
- Reference layers that depict demographics or other spatial attributes are available. If the defaults are not enough, you can search ArcGIS for more. I found that there are thousands of options that ArcGIS Online users have made publicly available.
- I could probably add a dozen items to this list — just try the maps yourself and use the existing Map or Filled Map alongside the new ArcGIS Maps.
Here’s a personal story showcasing how the new maps will transform how I use Power BI.
Earlier this year, I was asked to prepare a demo showcasing a sample “outage” map in Microsoft Power BI. If you have ever had a power outage, you may have checked your electrical utility’s website to see the extent of outages in your area and view their expected restoration times. For a standard example, here is my local energy company’s outage map. There was also a particular requirement to have outages cluster together as you zoomed out on the map. Armed with an idea, I built a quick sample dataset and went to work.
Could the standard Power BI map get me there? Somewhat, but not really. I could plot points. Clustering was not possible. This is what the customers for the potential client would see when they checked for outages:
The other map options for Power BI such as the Filled Map, Synoptic Panel and even the R visual (due to it being a static image) also could not satisfy all of the demo requirements.
As a result, I built a prototype using Leaflet and Leaflet’s marker cluster plugin. I found a GeoJSON file for a US city’s neighborhoods for added perspective and plotted my sample data. You can view the interactive map here or see a screenshot below.
I then started building a Power BI custom visual based on that Leaflet prototype, but…
Custom visuals are a lot of work… Too much work for a non-billable demo (and way too much work to get them to a state where they can be deployed in a production environment). I probably spent about an hour on the Leaflet prototype and two or three hours on adapting it as a demo-ready Power BI custom visual (clustering never worked by the way).
Fast forward to today, and the screenshot below shows the same data using ArcGIS Maps in Power BI. This took five minutes. That includes the time it took to find the Washington DC neighborhood outline reference layer from ArcGIS Online, build the custom DAX tooltip, and change the default colors and other formatting. This map also satisfies the original requirements with the ability to cluster points as you zoom out. With the light gray background, it also looks great and allows you to immediately focus on the data.
The main takeaway from this story for me is that I no longer need to waste time or tell someone “it’s not possible” to make functional and attractive maps in Power BI.
I’ll continue to use Leaflet and Mapbox for fun outside of Power BI. They have some great plugins that Power BI still cannot beat, but building custom maps is not my job — working with the Microsoft data platform is.
Between the Shape Map that was released earlier this summer and the new ArcGIS Maps for Power BI, I think Power BI has less need for time-consuming mapping workarounds. It took patience to get to this point, but Power BI’s mapping capabilities certainly took a step up.