Tableau vs. Power Pivot Part 11: Introductory Mapping

Today, we will talk about mapping in these tools.  Mapping is extremely important because it allows the user to see geographic patterns that may be missed if you were to use a more traditional chart type, such as a bar graph.  We will use a variety of sources in this post to illustrate the effective use of each of these types of charts.  For those of you that read our previous post on this topic, Tableau vs. Power Pivot Part 2: Basic Charting, you will remember that Power Pivot does not have any built-in mapping capability.  We also showed that Power View's mapping is somewhat lack-luster.  Therefore, we will introduce the final piece of Microsoft's "Power BI" toolkit, Power Map.  Power Map, formerly GeoFlow, is a new add-in for Excel 2013 that allows users to map their data in a variety of different ways.  The question remains, "How does it compare to Tableau?"  Let's find out.

Map 1: Choropleth

A Choropleth, a.k.a. Filled Map, is a map that is separated into distinct regions, where each region is colored according to a numeric value.  So, let's look at Profit by State, using our Superstore Sales sample data set from Tableau.  First up is Power Map.

Choropleth (Power Map)

As you can see, this map looks pretty good.  Power Map automatically connected to the underlying Power Pivot source, which could have been a big hassle.  Our big concern is that we had to look individually at 8 different background themes to get one that didn't overshadow the data.  On this same note, the map is filled with only one color, which make analyses on something like profit somewhat more difficult.  Let's see how Tableau does.

Choropleth (Tableau)

Tableau seems much cleaner to us.  The color scheme is much better for this case, primarily because Montana had a negative value, which we never saw in the Power Map version.  As far as ease of use goes, Tableau was simply a drag-and-drop experience, while Power Map forced us to look through a bunch of different styles before we found one that looked decent..  Therefore, we have to give this one to Tableau.

Winner: Tableau

Map 2: Heat Map

A Heat Map is a map that is not separated into distinct categories, yet is colored by how many values appear close to each location.  It is very good for finding "hot spots" among countable data.  So, let's look at John Snow's Cholera data.  For those of who have never heard of this map, it is arguably one of the most influential maps of all time.  Feel free to Google it if you are curious.  First, let's see how Power Map handles it.

Heat Map (Power Map)

This map looks really nice.  We can easily see where the deaths occurred and identify a couple of hot spots.  The default color scheme is also very nice as well.  This map did take a little bit of tinkering to get right, but it was well worth it.  Let's see what Tableau can offer up.

Heat Map? (Tableau)

To our knowledge, Tableau does have any built-in functionality for a geographic heat map.  The best we could come up with is the circle approach seen above, which really isn't any better than Power View could have done.  It still shows us the hot spots, but not as clearly as the Power Map version.  We also cannot easily distinguish when there are multiple sets of moderate values in close proximity.  This map wasn't too difficult to create, but it utterly fails in comparison to an actual heat map.

Winner: Power Map

Map 3: Journey Map

A "Journey Map" as we are calling is a map depicting the time-based travels of a person or thing.  It is great for seeing how efficient your shipping methods are or any other type of analyses surrounding that type of data.  In order to show this, we will use Napoleon's March to Moscow.  This is another very famous set of mapping data that depicts the catastrophic defeat Napoleon suffered when trying to march on Russia, originally created by Charles Joseph Minard.  We thought that this map would be great for examining the true mapping power of these tools.  This data, and the Tableau workbook used for this exercise, were originally compiled and created by Kim Rees of Information Aesthetics.  You can view her work on their website and on Tableau Public.  Now, let's see how Power Map handles this.

Journey Map (Power Map)

We tried to find any way at all of connecting these circles and we came up empty.  So, this is the best we could get from Power Map.  If we were inclined, we could replicate each of the rows in the data while slowly moving the locations so that we could fake a line.  However, that's quite a bit of work that most business analysts wouldn't have the knowledge or tools to pull off easily.  This map is ok.  It shows a general trend of the circles getting smaller as Napoleon marches to and from Russia.  Alas, the lack of a line pretty much ruins this graph's analytical capabilities.  Let's see what Tableau can do.

Journey Map (Tableau)

This map is so much more intuitive that the Power Map version.  We will concede that this map did take a little bit of Tableau knowledge to create.  However, the results are well worth it.  We can easily see how Napoleon was losing more and more troops as his journey went on.  Also, we removed the two branching routes from this view because they made the Power Map chart even harder to read.  This is what the original map looked like.

Napoleon's March on Moscow (Tableau)

It's pretty easy to see that Power Map is not very good at creating this type of map.

Winner: Tableau

Bonus!

Both Tableau and Power Map have the ability to create maps that change at regular time intervals.  These time intervals could be very short or quite long, it's up to the user.  Power Map's timing feature requires a datetime field to be present in your cube, whereas Tableau can work on any discrete field.  On the other hand, Tableau only has 3 speed settings while Power Map lets you explicitly define the duration of the show.  To be honest, we've never seen a situation where either of these features were put to good analytical use.  They are great features for "WOW"ing an audience, but not much more than that.  Also, Power Map has the ability to save virtual demos of your map as video files, which can be shared with other people.  Nevertheless, we chose not to let these features impact our experiment.

Also, an anonymous reader who works for a Canadian company pointed out that Tableau doesn't support the names of Cities/States/Provinces, etc. outside of the United States.  Unfortunately, we cannot confirm nor deny this comment.  If you know of any specifics to this situation, please let us know in the comments.

EDIT:  A different anonymous reader posted this link that shows Tableau saying that they support many different regions/states/etc. outside of the United States.  If you find that your cities is small enough that it is not in Tableau's database, then you should easily be able to geocode your city into the data file.

It should be noted that Tableau uses a static geocoding source, whereas Power Map can pull it's geocoding directly off of Bing Maps.  However, Tableau does allow the user to create custom geographic dimensions.  You can supply Tableau with the dimension values, latitudes and longitudes.  Then, you can use that geographic dimension on any chart you want.

EDIT:  A curious commentor questioned whether or not a security risk is posed when Power Map queries Bing Maps for the mapping information.  Is any of the data left vulnerable?  Could a clever snooper intercept this?  Let us know in the comments.

Summary

We saw that Tableau beat Power Map in its ability to create Filled Maps because of it's cleaner aesthetics.  However, Power Map fought back by creating a truly inspiring Heat Map that Tableau was unable to replicate.  Finally, Tableau showed its true colors by creating a Journey Map that Power Map could not compete with.  Thanks for reading.  We hope you found this informative.

Brad Llewellyn

Associate Data Analytics Consultant

Mariner, LLC

brad.llewellyn@mariner-usa.com

http://www.linkedin.com/in/bradllewellyn
http://breaking-bi.blogspot.com