Tableau Reviews

I spend most of my time with clients trying to get more out of their existing tools, more often than not some combination of an ERP tool (JDEdwards, SAP etc.) a selection of more “pure” databases such as Oracle, Access, SQL and a lot of spreadsheets.

All of these tools have well understood strengths and weaknesses. One common issue is the effort required for data visualisation. Just to be clear, when I say data visualisation, I mean the presentation of data in a graphic format so that trends, insights and results can be easily and quickly extracted by the reader.

Excel is fine, especially as most of us are very familiar with it, but it has definite limitations. Try putting 10,000 data points on a scatter plot or add interactivity to your chart and you’ll see what I mean.

I’ve taken a look at three tools that really caught my eye when looking for a client solution. There are 20 players in the 2011 Gartner review of BI tools. Many of these are 400lb gorilla solutions, like Cognos, some are more sector specific and others tend to be strongest in certain geographic areas.

Being a Brit, who wanted something strong on visualisation, quick implementation and didn’t want to build a data warehouse, the shortlist came down to:

• Tableau
• Qlickview
• LogiXML

In this blog I’m going to look at Tableau. Over the next few days I’m going to add short reviews of Qlickview and LogiXML.

I’m not Gartner, so there may be some other contenders out there, but life is short so it’s these three that I’ll be testing over the next few blog entries.

Just to set expectation, I’m a data professional, but I’m not a coder or database specialist. I come to this as an “IT competent” management information specialist. I know the sort of output I want to see but I don’t like reading manuals or spending ages building things – much like my clients.

Installation

Installation of Tableau is pretty straightforward. You need to register on their site to get a two week trial period. You download a 51MB .exe file, double click it and in a few minutes you have a working installation of Tableau. It’s a Windows only affair. For this review I ran it as a virtual installation on a Mac which didn’t seem to cause any issues

First off the program asks you to connect to a data source. For ease I used one of the ready-made databases, but the connectors on offer look pretty comprehensive.

Tableau asks you to connect to a data source, including Excel of course

Interface

The interface is fairly busy, with a number of different areas on the interface. It’s all quite clean, logical and modern. There’s a tabbed workspace concept which anyone from an Excel background with instantly be comfortable with.

The chart creation interface will be very straight forward to anyone who has used Excel pivot tables

Applying the “man test”

Like many slightly vain middle aged men, I’m not too keen on tutorials and manuals. I dived into Tableau without reading any of the documentation. Within 15 minutes I’d figured out how to produce and tweak some pretty good looking graphs. I used the sample data provided by Tableau, which may have helped things along a bit.

Flashy geographic mashups are a doddle

This dull looking bar chart has hidden depths - you can click to drill down. It didn't require any configuration

Charts like this heatmap are quick and simple to create

To create a chart you just drag and drop the data items from the panel on the left into a simple table. It’s a lot like creating a pivot table in Excel, only a bit less impenetrable.

Another 10 minutes and I’d figured out how to produce a quite credible looking dashboard. So in terms of “ease of use” I’d give it 10 out of 10. I think it goes as far as you can with usability before you need to start throwing features overboard to simplify it further.

Not a great design by me, but it is a very quick and simple job to create this live dashboard

Features that really stood out

Creating charts: It’s so clearly the “sweet-spot” of this package that it’s easy to overlook the ease with which charts are created.

Drill down: It’s trivially easy to rollup or drilldown through data in the appropriate graph type

Unusual charts: Tableau can create charts that are either difficult or impossible to create in Excel. Heat-maps, graphical mashups (see this dental decay graphic one for a good example) and interactive charts are all possible. You can then post these on the inter/intra net for others to view.

Tableau was dismissed by one of its rivals salesmen as having “pretty graphs”. After a few hours with Tableau I think it’s fair to say that it has some of the best conceived “template” graphs I’ve seen in any package. Their blurb on their web site talks about “visual best practices are baked-in“ and they aren’t kidding. There’s minimal design clutter, clear layout and very spare use of colour. No 3d bars, shading or gridline clutter here. The type of design approach championed by Stephen Few has rarely been executed so well and consistently.

Sharing

The samples on the Tableau web site shows a good selection of ways you can share:

As a web page – you can drop a javascript code snippet into a page to create a web page or share a link

As an image, data download, crosstab or print to PDF

A download to the free Reader application

Samples

To get a good impression of what Tableau can do it’s worth having a look at their excellent selection of demo graphs. Although it’s not always obvious what purpose these graphs serve, there are some outstanding examples of rich, intelligent chart design. I particularly like this dental decay graphic and this flight delay graphic.

The key thing for me is that I actually enjoyed using this software and started working out whether I could justify the cost of a personal copy. Pricing is $999 for a desktop copy that can’t access databases, $1999 for the pro version that can plug into just about any data source – so for personal use it’s not cheap. As with most vendors, there’s no mention of the client-server version software costs, but it’s likely to be industrial grade pricing, like their competitors.

Where I see Tableau (and data exploration tools in general) really delivering value is in encouraging managers and analysts to take the terabytes of data that already exists in most organisations and spin it in new and insightful ways. So rather than just looking at the classic complaint volumes and complaint ages (for example) bar charts, why not start looking at complaint severity and customer segment, or complaint source by geographic source within the production area – as a visual heatmap' Many organisations have allowed their tools to define their ambitions for their analysis, with a tool like Tableau it’s an opportunity to be a bit more ambitious.

I’d recommend downloading a trial and giving it a spin.

Note: I have no commercial relationship with Tableau, so my view is as independent as they come.

What is our primary use case?

What is most valuable?

What needs improvement?

For how long have I used the solution?

What do I think about the stability of the solution?

What do I think about the scalability of the solution?

Which solution did I use previously and why did I switch?

How was the initial setup?

What other advice do I have?

Which deployment model are you using for this solution?

If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?

What is our primary use case?

What is most valuable?

What needs improvement?

For how long have I used the solution?

What do I think about the stability of the solution?

What do I think about the scalability of the solution?

How are customer service and technical support?

Which solution did I use previously and why did I switch?

How was the initial setup?

What about the implementation team?

What's my experience with pricing, setup cost, and licensing?

What other advice do I have?

Which deployment model are you using for this solution?

What is our primary use case?

How has it helped my organization?

What is most valuable?

What needs improvement?

For how long have I used the solution?

What do I think about the stability of the solution?

How are customer service and technical support?

How was the initial setup?

What about the implementation team?

What's my experience with pricing, setup cost, and licensing?

What other advice do I have?

Which deployment model are you using for this solution?

What is our primary use case?

How has it helped my organization?

What is most valuable?

What needs improvement?

For how long have I used the solution?

***We've upgraded to 2018.1 So I've updated the Pros and Cons accordingly!"

I've been working between Tableau and Qlikview (depending on use case) and have been jotting down Pros and Cons of each as I come across them.  Many of these, but not all, are centered around some fairly specific UI functionality and limitations that I've come across while trying to deliver on various dashboard solutions.  If you're evaluating Tableau as a potential data visualization solution, some of these points might help with your evaluation, or at least I'm hoping so!  One caveat I will apply to the "Cons" section is some of these may in fact be resolved in the latest version, or may be achieved through some means that I'm just not aware of (though I do generally research pretty extensively before throwing in the towel).  

Pros/Strengths

• Multi-dimension Graphs, side by side bars
• Formulas/Calculations are a familiar structure to many  (IF..THEN.. ELSE)
• Visualization attributes:  Marks – Color, Size, Label, etc.. Easily Accessible and Intuitive
• Very good Geo/Mapping capabilities
• Built in Table Calculations (% of Total, Rank, etc..) Update:  Even more intuitive now.  
• Publish to PDF.  Despite trying to encourage users to interact with the live tool, there seems to always be some need for this. 
• Story Creation for presentations.  
• Free Tableau Reader – You can create ‘Packaged Workbooks’ and those with Reader can open the file and manipulate the dashboard.  (No ability to refresh the data)
• Easy Cube Connection.  Though it can still be quirky and limited relative to using other data sources.  You can't for example, use LOD calculations.  If you plan to use Cubes as a data source, be sure you thoroughly understand the limitations.  
• Tool Tips (Hovers).  Easy to add all kinds of additional data to hovers.  Update:  The Viz in tooltip feature is fully baked and in certain circumstances, very useful!   
• Adaptive sizing based on display resolution.  Also something Qlikview does not do - you must develop for particular display size/dimensions.  Update: I've since learned that it generally makes life easier if you still used fixed dimensions for your dashboard.
• Drag and Drop Hierarchy creation
• Device Layouts (Mobile, Tablet, etc.).  Update: Rumor has it Tableau is currently working on a better Mobile solution.
• URL Actions - Allow you to enable workflow between Tableau and other Applications through click-throughs and parameter insertion in URLs.  

Cons

• Small multiples (a.k.a. Trellis charts) are possible only through very hacky means.  Update:  Still remains a challeng
• Sparkline Tables are not supported.  There are tutorials on how to accomplish this, but as with small multiples, it is hacky and limited. Update: Still remains a challenge.
• Heavy data prep needed.  Blending within the tool is clunky and causes performance hits.  Some functions are limited or unavailable when using blended data (e.g. LODs) Update:  Blending has improved but it is still generally a better practice to do all your data prep prior to ingestion.  Tableau Prep now exists to help with this, but it is still very much in its infancy.
• Sorting on joined data requires hacks.  Update:  Unchanged
• No easy ‘Clear’ to remove applied filters and/or Exclusions/Keep Only, which makes it sometimes easy to lose sight of how exactly the data has been filtered. Update:  Unchanged
• Combo charts unavailable without Date dimensions  Update:  Unchanged
• Dynamic column headers/aliases difficult and hacky.  I find this limitation particularly frustrating.  Example use case:  I have two columns - Current Month, & Previous Month.  As these will always be dynamic, I want them instead to display the actual Month as it changes (e.g. Apr-2016, Mar-2016).  Update:  Very unfortunately Unchanged.  
• Can’t format a single measure column in a table (e.g. make one bold or conditional format)  Update:  I've heard rumors that this is now easier, but have yet to identify how.
• Image thumbnails in tables not supported.  In my last two roles this has been a fairly large impediment.  I'm likely not the only one.  Qlikview, SSRS, & Even MS Powerview/PowerBI are able to support this. Update:  Unchanged 
• Large Table reports with many filters (our case was 14 columns, 300K rows, and about 8 quick filters)   to narrow the data set – Don’t bother!  It will be excruciatingly slow even on the server. Tableau will likely say this is not what it is intended for anyway (it's a Data Viz tool, not a Report tool), which is fair, but sometimes it's necessary as part of a larger solution. Update:  This seems to have improved some though you still run into the limitation of 16 columns.  Rumor has it, this will be done away with in future releases.
• 
  
Any questions, or comments?  All feedback is welcome.

What is our primary use case?

How has it helped my organization?

What is most valuable?

What needs improvement?

For how long have I used the solution?

What's my experience with pricing, setup cost, and licensing?

Which other solutions did I evaluate?

What other advice do I have?

What is most valuable?

How has it helped my organization?

What needs improvement?

For how long have I used the solution?

What was my experience with deployment of the solution?

How are customer service and technical support?

Which solution did I use previously and why did I switch?

How was the initial setup?

What about the implementation team?

What's my experience with pricing, setup cost, and licensing?

What other advice do I have?