5 Ways to Create Maps in Microsoft Power BI

UPDATED June 3, 2017: There have been a year’s worth of updates to Power BI since this “5 Ways” post was originally published. Please see the updated “10 Ways to Create Maps in Microsoft Power BI” instead.

-David

***************************

Now there are 7 Ways to Create Maps in Microsoft Power BI…

UPDATED September 26, 2016:

Announced today, there will soon be Esri ArcGIS maps for Power BI – at no additional cost. In my mind, its eventual release will make Power BI instantly relevant as a “self-service” as well as basic enterprise GIS tool. There has always been a noticeable gap in functionality between Power BI and Tableau (both Tableau native and with Mapbox and WMS integration). With Esri ArcGIS maps, I believe that gap will be closed.

Read more about the new capabilities coming soon to Power BI with Esri maps: http://www.esri.com/software/arcgis/arcgis-maps-for-power-bi

My commentary on the BlueGranite blog: https://www.blue-granite.com/blog/power-bi-arcgis-maps-closes-a-critical-gap

UPDATED August 2, 2016:

As of the July Power BI Desktop update, the new Shape Map exists as an additional method. The Shape Map allows users to select from a handful of provided geographies as well as load a TopoJSON file for custom shapes. You can read more about the Shape Map at the following locations:

• Documentation: https://powerbi.microsoft.com/en-us/documentation/powerbi-desktop-shape-map/
• Commentary: http://www.kasperonbi.com/use-any-map-with-power-bi/
• TopoJSON file collection: https://github.com/deldersveld/topojson
• Troubleshooting: https://dataveld.wordpress.com/2016/09/01/overcoming-potential-power-bi-shape-map-rendering-issues/

ORIGINAL POST:

Location data is nothing new, but it seems that people are storing more of it. Regardless of the reasons for this perceived increase, the data is there. For users that want to visualize their geospatial data on a map, there are numerous choices to make. Sometimes, these choices are constrained or dictated by technology, but in other cases, technology allows for greater flexibility. When it comes to Microsoft Power BI, what are the options?

Before getting into maps in Microsoft Power BI, it is also valid to state that a map is not required to represent geographic data. A large temptation is to throw all location data onto a map, but sometimes a list of locations and a bar chart or another simple representation will do. In many cases, forcing data onto a map can be more confusing or potentially misleading than using another type of visualization.

While there are numerous options to create maps in Power BI, there is no method that stands out as “the one” that is best for every situation. Each method could have advantages in different circumstances. The following list is not a ranking, but I have included a small sample of pros and cons for each method. Feel free to comment and build up the pros and cons, or correct anything that may stand out in error.

1. Native – Map & Filled Map

Power BI has two native map visuals: the map and the filled map. The map displays points, which can optionally be sized as area bubbles. The filled map displays shaded geographic areas. These are good maps for standardized display but fall short when it comes to more advanced customization.

• Pros
  • Accessible, easy to use, and fully supported by the Microsoft Power BI team
  • Flexible in that it can geocode location data addresses or work with stored latitude and longitude coordinates
  • Good basic maps for common uses
• Cons
  • Smallest available / default point size is too large, which leads to too much overlap in locations with high point density
    • UPDATED April 2017: Within the past few months, Microsoft created a Format option for “Bubble” size. Points can now appear much smaller than they originally could.
  • Base tiles – no ability to change the richly colored HERE tiles, which reduces the color selections for data and may make points more difficult to distinguish from the background
  • Occasional Bing geocoding issues where points or filled polygons end up in the wrong location (particularly noticeable when using County level data)
  • No current support for standard geographic formats or custom layers (GeoJSON, ESRI Shapefiles, etc.)

2. Custom Visual – Synoptic Panel

The Synoptic Panel can be found in the Custom Visual Gallery and has a companion website (Synoptic.Design) that contains an editor and templates. For the latest version, always check the Synoptic Design website rather than rely on Power BI’s Custom Visual Gallery. While the Synoptic Panel allows for a broad and powerful range of different visualizations, it can be used for cartography. For example, my former BlueGranite colleague Jason Thomas (b/t) recently published a post detailing his use of the Synoptic Panel for a hex tile map of the United States.

• Pros
  • Some templates available for common geographies in the Synoptic gallery
  • Fully customized areas for any image
• Cons
  • No support for standard geographic formats, i.e. cannot import an existing GeoJSON file
  • Time-intensive compared to other Power BI map options (but end result is worth the effort)

3. Custom Visual – Globe Map

Another option from the Custom Visual Gallery is the Globe Map.

• Pros
  • Whoever created this employed ThreeJS in a Power BI visual, which is a solid technical achievement
  • Provides small level of continuity for anyone who used Power Map for Excel (and does not mind a severely reduced feature set)
• Cons
  • 3D vantage point makes it difficult to simultaneously view and compare areas across the globe and may also cause perception issues
  • Has difficulty clearly rendering base tiles at low-level zoom
  • Lack of formatting options for bar and heatmap

4. Create your own Custom Visual

For anyone who is truly ambitious, they can create their own custom visualization for Power BI. I went down this path to some extent last autumn, and I know of a separate Leaflet visual from James Dales (t) that has not been publicly released. If any of the prior methods to map data in Power BI did not suit you, and if you have the time and expertise, a custom visual can be a great way to display your geographic data.

• Pros
  • Extensive control over output
  • Can incorporate custom geography, GeoJSON, Shapefiles, etc.
• Cons
  • The most time-intensive option of all
  • Difficult to use for anyone unfamiliar with TypeScript and JavaScript visualization libraries such as D3 or mapping libraries like Leaflet or OpenLayers

5. R Visual

For anyone who wants to go beyond the basics, Power BI also provides the ability to create visualizations with R.

• Pros
  • Extensive and detailed control over output
  • Can incorporate custom geography, GeoJSON, Shapefiles, etc.
• Cons
  • R visuals are currently only available in Power BI Desktop and not the online Service
  • R visuals require static images as output, which does not take advantage of R’s full capabilities
  • Difficult to use for anyone unfamiliar with R

I have intentionally left out methods such as plotting latitude and longitude on a scatterplot with or without an image background, or using SandDance. Alternatives like that can sometimes work, but distortion often comes into play, or it requires a large number of points to make the geography recognizable.

Also, I have not included known future functionality such as native support for custom geography. Based on confirmation from the Power BI team that is now stated publicly on on the Power BI Ideas site, a new “Shape Map” will be added to Power BI soon that will include at least GeoJSON and TopoJSON data.