Tableau vs Power BI

I’ve had the fortune in the last couple months to be doing a lot of innovative work for a client of mine using the power BI toolset. Recently, I’ve been busy publishing information with tableau, and have come to realize the strategical difference in each toolset. Microsoft’s direction is now devices and services, which is why they are pushing their cloud solution.  They aren’t alone, if you have a look at TIBCO’s spotfire, they are up to the same thing.  Even tableau has begun to offer some on prem / off prem services.

In the BI world, the cloud analytics world becomes only a “anywhere anytime” presentation layer that enables ultimate business flexibility. That said, the underlying DAX platform and powerpivot tabular model empower the data analyst to further manipulate the data. In the analytics space, quick iterations are the key to success. One needs to be able to iterate quickly to run down the path of data discovery. Quicker iterations lead to hasty data cleanup and more insight sooner. Powerpivot is Microsoft’s latent answer to a true semantic layer in an analytical tool. It’s built on the same engine as SSAS, only that it’s a tabular model, and not multidimensional. That said, most organizational needs can be accomplished quickly (and with more iterations) with a simple tabular model. No need to put the cart before the horse and build a complex model before you know what too look for.  Powerpivot gives a good visual representation of what the tabular model looks like…

powerpivot model

The following is an example of what the Data Analysis Expressions (DAX) looks like.  It’s not meant to be a full semantic language, but rather a more efficient replacement for excel formulas and calculations.  As always, it presents an opportunity to bend the model when you need it to work…


What truly differentiates powerpivot and the tabular model is the ability to pull in data from any type of service (OData, Excel, MSSQL, Oracle, you name it), but the ability to manipulate and transform that data in the model itself.  This is what differentiates this product from it’s predecessor and it’s competition.  However, Microsoft doesn’t allow these models or this analysis to be published publicly… that’s where tableau comes in.

In preparation for the YYC hackathon, I’ve been looking at platforms that I can publish data publicly.  The choice ended up being either tableau or D3.  D3 is a data presentation layer based on javascript, which works well for displaying information, but that platform doesn’t help an analyst iterate quickly.  I opted instead to sign up for tableau public, to see what it can do.  The public version only allows the user to import flat files, while the desktop version (free trial, $1k licensed) allows the user to import from a variety of data sources.  That said, the public version allows the user to do basic joins, and filtering, but allows for little flexibility in manipulating data on the fly.  It seems to me, that any complex data manipulation must be done before it’s brought in the tool.  Given that assessment, data manipulation can be done in any database, and exported to a flat file where tableau can pick it up.  The only value tableau desktop offers, is the ability to do some of that complex logic within the database itself.  Otherwise, I might as well just import all data sets into a tool, and export them back to a flat file for tableau pickup.

So that is exactly what I had to do for the hackathon.  Import the data into my own database, manipulate it, and export it as a series of excel files.

Screen Shot 2015-04-18 at 4.38.36 PM (2)

Messy, but it worked.  Tableau allows for some basic manipulation (you can join files, filter them, the basics etc).  In effect, we used the tableau public service as a presentation layer.  But that’s just it, tableau only offers a presentation layer.  It’s semantic layer isn’t very sophisticated.  The presentation layer is more powerful than the powerBI visuals, but it lacks in presentation, it makes up for in flexibility and efficiency.

Analytics tools exist in a commoditized space.  As Bill Ruh said last month, Efficiency is the new competitive differentiator.  He’s right.  The more efficiently data scientists can mash up with a tool, the more valuable it becomes to them….

At least Barend and I were able to leverage Tableau for the hackathon.  Good news,  won third place!


Tags :
Far better an approximate answer to the right question, which is often vague, than the exact answer to the wrong question, which can always be made precise. -John Tukey
The plural of anecdote is not data. - John Myles White

Recent Posts

RSS PowerBI blog