It’s Time to Reevaluate the Power BI “Map” Visual

It’s Time to Reevaluate the Power BI “Map” Visual
Reading Time: 3 minutes

One of the first things I tried in Power BI Desktop (née Designer) in late 2014 and early 2015 was the Map visual. Power BI’s mapping capabilities were spartan compared to other tools I was using at the time such as Excel, Reporting Services, Tableau, and Leaflet.

I wanted Power BI to do everything that I could do in Leaflet. Clustering, minimap, heatmap, custom feature layers–you name it. Because of my perceived early challenges getting the maps I wanted, a Leaflet map was the first thing I attempted as a custom visual back in September 2015. I didn’t realize at the time how much my personal views differed from the average Power BI consumer.

After a few years and multiple projects with Power BI, I’ve changed my mindset. Power BI is never going to be a high-end geospatial tool, and that’s okay. It’s an end to end business intelligence tool.

That doesn’t mean things have been stagnant since 2015. Over time, I’ve taken an interest in many of the custom visuals that people have published. I listed several pros and cons for various map visuals in Power BI in 2017 with my 10 Ways to Create Map in Power BI post. Since that time, there have been several improvements to the core map visuals as well as to numerous custom visuals. We’ve come a long way since those early days.

Map Visual Improvements

Bringing a more specific focus to this post, the March 2019 release of Power BI Desktop has several additions to the core Map visual. If you’ve previously minimized the use of the Map visual or have settled for alternative custom visuals, it may be time to reevaluate the Map in light of newer enhancements.

1. Smaller Point / Bubble Size

One of the primary drawbacks of the Map visual prior to March 2019’s update was that the point / bubble size was large even when set to the lowest option of 0. The lowest setting is now -30. While a negative setting is a little strange at first, it makes sense if the product team didn’t want to impact existing reports by having every map suddenly change point sizes.

Consider the image above compared to how things looked before March’s update. Hopefully you can see why I’m so positive about the negative. This change in itself will tempt me to select the Map visual in basic scenarios instead of alternatives like the ArcGIS Map or a custom visual.

2. Heat map

The Power BI Map now includes a Heat map option. Changing the visual from points to a heat map provides a better visual representation of density. One drawback of the Map when there is highly dense data is overlapping points of the same color (and remember–these points were HUGE until March 2019 even with Size set to 0). A heat map shifts the visual so that color directs attention toward areas of high or low density.

In addition, one of the nice options is the ability to switch units between Pixels and Meters. This gives designers control over how things display at different zoom levels.

3. Grayscale and Dark Themes

The Grayscale and Dark themes have been around for a year, but I haven’t updated my 10 Ways post since those options were released. Back in the day, the Road theme was the only option. From a data visualization standpoint, using Road isn’t ideal because its many colors take the focus away from the data plotted on the map. The simpler Grayscale and Dark themes shift a viewer’s focus to the data and also allow for a wider choice of data colors that will not blend as easily into the background map.

Honorable Mention

In addition, the Map also includes optional Zoom buttons. This isn’t a major feature on the same level as the three improvements above, but it makes zooming more intuitive for users by having the control visible on the map itself.


Kudos to the Power BI product team for improving the Map in so many ways recently. With the new additions to the Map in March 2019, I’m going to use this visual more often. No, it’s not going to take the place of more advanced capabilities. I’ll still rely on specific features in the ArcGIS Map, Mapbox Map, Shape Map, Synoptic Panel, and Icon Map in particular. For basic business reporting though, when I only need to plot data on the map, the Map will do.



  1. It’s good update for the native map visual. However, looking to embed multiple layers to map visual dynamically as well custom icons is still pending and differentiates Tableau and Qilk from Power BI. Tableau and Qilk allows you to integrate multiple layers in a single map visual.

    It’s bit challenging to update underlying layer in ArcGIS Map dynamically if you would like to work with your custom layer. There is still manual work involved publishing layers in ArcGIS Online.

    I hope Microsoft is listening this and come up with some serious mapping features with native map visual or create Azure Maps Visual with advanced capabilities.

    1. “Power BI is never going to be a high-end geospatial tool.”

      I’ve devoted the past 13-14 months of my life to writing a Power BI Layered Map tool (taking precious time off to write this post), I have some thoughts on the matter. Power BI is still new to teams that use Tableau / Qlik / SpotFire and I think it will take a few more years before Power BI firmly replaces them, but that day is definitely coming, the capabilities and cost of Power BI is going to dominate the space. Tableau and SpotFire have a university presence that Power BI does not (but should). Microsoft has weak mapping tools to build a powerful map widget from compared to MapBox, ESRI, OpenLayers, etc. I think they will need to make an acquisition to change that (and probably wont?).

      I work for an energy company where mapping is very important – and the need for a best-in-class interactive map with heatmaps, clusters, visual-join-to-power-bi-data, streaming WFS, popups, images, labels, drilldown, feature-navigation is a firm need.

      The complexity of this tool is massive. Determining the behavior of what should happen when users configure a heat-map enabled power bi point layer with clustering, popups, selection, and this layer is joined to other feature layers and some of those feature layers are streaming through WFS – the end users would need to grant the developer huge latitude for bugs in situations like this.

  2. Pingback: - DataVeld

Leave a Reply

%d bloggers like this: