Tableau Request Live – The 2nd Episode: Data Densification

TRL

For this week’s Tableau Request Live episode (a Saturday exploration of Tableau with Joe Mako and yours truly, recorded live), we dig deeper into the complexities of Data Densification in Tableau.  The 1st Episode of TRL focused on a specific Tableau problem of solving a User Retention Rate with a Single Date and ID field, and we briefly touched on some of the concepts that are explored in more depth in the 2nd Episode.

No, “Densification” is not a term you’ll find in the available Tableau KB articles or other official documentation available from Tableau Software.  Data Densification is an umbrella term power users have used to describe scenarios when marks (data) are added to your view.  There are two main types: Domain Completion and Domain Padding.  Both types of Densification behaviors are explored in the video, and yes, there are cases when both occur at the same time.

In brief, Domain Completion occurs when marks (data) are added for potential Dimension value combinations in a view/worksheet.  Domain Padding is what happens when marks (data) are added for potential values in the range of a Range Aware pill, which includes both Date Dimensions and Bins in Tableau.  This is typically “turned on” by using the “Show Missing Values” option on a Range Aware pill, which makes it a bit easier to define and understand, from my perspective.

I could try to write my own synopsis of the video itself, but the recording and the resources linked within the video description are meant to serve as an ongoing reference point — what is shown in the video and written in the description and links is based on the in-depth study and current working knowledge of many Tableau users, including (but not limited to) Joe Mako, Jonathan Drummey, Richard Leeke, and Jim Wahl (there are many others who have contributed to this ongoing understanding, including Tableau employees, developers, and many other Tableau users in the Community).

Again, the documentation and content of the video itself are simply based on the current working knowledge of Tableau users in the larger online community, and the current version of Tableau software (8.2 as of the day the video was recorded and this blog post was written).  We hope the video becomes obsolete very quickly, and that we will have more direct control over Data Densification in future versions of Tableau.

With that being said, let this video serve as one reference point for Data Densification behaviors in Tableau. In the live recording, Joe and I walk through the two main types of Data Densification (Domain Completion and Domain Padding), conditions to turn it on, how to turn it off, and exceptions. Some of the scenarios shown in the video are edge cases, but if you have ever wondered why Tableau is making more marks then there are records, Data Densification is the reason.

The awareness that Densification may or may not be occurring is what we hope users will gain from resources like this.

Joe and I hope you can gain an awareness of what is happening when your number of marks suddenly changes on your Tableau canvas!  Sometimes Data Densification behaviors are helpful and they may allow us to build views that we otherwise could not — we would simply like to have more direct control over whether it is ON or OFF in a given view.


I, personally, hope other users see the videos we are making, and start making their own!  I’ll say this, as well: Joe Mako spends hours, every day, helping Tableau users (and I consider him to be the “Teacher” in our relationship) — reach out to Joe, or someone else, and teach/learn collaboratively!  These videos are recorded using a free online application, http://www.freeconferencecall.com/ and it is VERY easy to record a video like this one.


Below are some further resources on the topics explored and discussed in the video:

Document used in video:
‘Tableau Data Densification’

 

Jonathan Drummey’s post mentioned in the Video:
‘At the Level – Unlocking the Mystery Part 1: Ordinal Calcs’

 

Two related threads from the Tableau Community Forum:
‘Can I add rows to underlying data from Tableau?’: http://community.tableausoftware.com/message/299456
‘Historical Chart on WINDOW_AVG data not working’: http://community.tableausoftware.com/message/299403

 

Keith Helfrich‘s “Master Tableau Concepts”:

 

Tableau Software’s Customer-Generated “Think Data Thursday” Video Library:
(videos on Domain Padding, Data Scaffolding, and Table Calc videos are all related to Densification in one way or another)

 

Another plug for Jonathan Drummey’s Blog, which has several posts that discuss related concepts:

Until next time.

Advertisements

Tableau Request Live – The 1st Episode: User Retention Problem

TRL

You remember this, don’t you?

DISCLAIMER: The content of the video linked below gets fairly complex. While the video may be rough around the edges, I hope it serves as a resource for building an awareness of what Tableau is doing as you build a viz. This blog post is not meant to serve as single source of knowledge, but just one of many resources that may help you as you encounter unexpected results inside Tableau.

Overall Goal

The overall goal of this video response was to show some of the complexities in arriving at a seemingly simple solution inside Tableau–in this example, we wanted to show a single line which represents the User Retention Rate along a Date — the granularity of the final line chart view is at the level of detail of Date (so we want 1 Mark per Date).

However, as you will see in the video, the problem encounters some “issues” along the way – Joe resorts to Nested Table Calculations, with some advanced concepts sprinkled in throughout (use of the Path Shelf with a line chart, filtering marks down to only what is needed to build the final result, and nesting Table Calcs within other Table Calcs, just to name a few)…  for more on Nested Table Calcs, and some of the other more difficult concepts to grasp in Tableau, I would recommend the TDT Video Library as one good resource where you can learn more from other amazing Tableau users.  I should note that I am not an expert on all things Tableau, and I would benefit from reviewing those videos just as much as anyone else!

I got involved with Joe, because I understand a lot of the Table Calculations he used in this solution — what I lack is an understanding of how all the pieces fit together, to work toward his solution.  For my part, I had some personal goals, in addition to showing how to work around problems in Tableau.  They included: picking up on Joe’s skills/knowledge/understanding, participating in a collaborative learning environment, and spreading knowledge of Tableau’s complexities to others.

The Live Video Recording

In a short 44 minutes, Joe Mako and yours truly recorded a live video response to a Tableau Forum‘s question on User Retention.  The original Forum post is here: http://community.tableausoftware.com/thread/148852

I emailed Joe on a Saturday afternoon, and within a few short hours, we were set up to record this live video response.  Thank you Joe!  You are amazing, and oh so generous with your time!  I would encourage other users to reach out to people like Joe, as I have found them to be very encouraging, generous, and outright helpful in many situations.

Now, the original Forum question is about calculating a User’s Retention Rate along a date.  Joe did simplify the “data” down to only two necessary dimensions for the problem at hand: an ID, and a Date.  Building on that simple dataset, he is able to arrive at a solution in short time.  However, there are many Tableau-specific issues that are encountered and worked-around during his workflow.

As stated, you will see some sophisticated Tableau concepts being used and described in this video.  It may difficult to follow if you do not have at least a basic understanding of Table Calculations in Tableau.  As a starting point, you might look at this list of resources from Jonathan Drummey, if you aspire to learn more about Table Calculations: http://drawingwithnumbers.artisart.org/want-to-learn-table-calculations/

The Video Response, which I’ve suggested is something akin to ‘Tableau Request Live’ (similar to the popular MTV show from your youth – or possibly, your children’s youth…), is hosted at the Vimeo link below:

Closing

Let this serve as evidence that no one Tableau Zen Master is perfect, as all of us make mistakes inside Tableau, and it can be very difficult to keep track of the many different factors at play in any given situation. It takes practice and study to understand what Tableau is doing, especially once we dive into more advanced areas of Tableau, like Table Calculations — as Joe will tell you, there is no documentation or video in the Tableau Trainings series to show you about Data Densification, let alone densification specific to using the Line mark type.

We have to build toward a solution inside Tableau, and I am reaching out to others like Joe, to learn how to better connect the dots in my own knowledge and understanding…to increase my awareness of how Tableau performs its amazingly fast rendering of marks on the canvas.  Reaching a solution for this problem would most likely take me a lot longer than it did Joe.  So, I’m partly involved to learn more myself — I hope to add commentary and questions as we work through some problems together, in order to further serve the Tableau community as a whole..

And, remember, we spent a small chunk of our Saturday doing this; I’ve tried and tried to learn this stuff throughout my work day, but find it easier to focus when I can set aside time outside of work hours.  This is just a personal comment on spending time outside of the daily grind with Tableau — without doing that, there is no way I would have been named a Tableau Zen Master for 2014. Also, I would never have been able to collaborate with Joe, if I had not asked. Sometimes, just asking others for direct help is the biggest step to overcome.

Finally: As described in the video, we certainly hope a video like this will become obsolete soon.  We dream of a day where building this type of solution inside Tableau is far less intimidating – a day where a Tableau user can continue in their workflow without being interrupted by unexpected results.

Until next time.

Tableau’s TC14 Conference – #DATA14 in 4 Parts

Well, I’ve just returned to Indianapolis and my first priority was to get a Tableau-centric blog up and running, so here I am!

As a first post, and to get the ball rolling, I wanted to share four main reflections on this year’s Tableau Conference (#DATA14):

More is Coming

Whether you watched the TC14 Keynote or not, many things are coming down the pipeline:  Performance improvements seem to be a large part of this, but I also saw some interesting visual “helpers” in upcoming releases–things like highlighting marks while using Table Calculations, so the user can more easily understand concepts like “Partitioning”, “Addressing”, and “Compute Using” (nothing in the demo about “At the Level” unfortunately).  A tablet-specific Server app, quick wins in Formatting Story Points, more visual options for using and editing calculations, reference lines, and a many, many more great features were announced.  You can watch the Keynote to get all the details on what was shown in the Demo portion.  I hope to keep this post short and sweet…  like the Conference itself, there is no way I will cover everything I would like, so I’m just sort-of winging it here!

The User Community is Amazing… and Growing!

I thought I knew the “Community” of Tableau Users from being active on the Tableau Forums, but I quickly learned I know very few of the users who interact with Tableau on a daily basis. Being named a 2014 Zen Master certainly helped me meet some power users, but I also met many everyday users who struggle–and have great successes–with Tableau on a regular basis.  It was amazing to meet all types of users.  Special shout outs to Tableau users/customers Richard Leeke, Toby Erkson, Keith Helfrich, Shawn Wallwork, Noah Salvaterra, Jonathan Drummey (whose blog, Drawing with Numbers, has been a great source of learning and inspiration over the past year) and many, many more for making me feel welcome and for teaching me a thing or two in the process.  This was my second conference, but as a newly appointed Zen Master, it was quite different from my first. It was also the first time I have presented at the Tableau Conference–my session is listed here.

The opportunities were vast, and  the fun never ended. I’m still buzzin’ with thoughts in how to apply what I saw or heard in sessions…

Tableau Rocks a Conference

Sir-Mix-a-Lot, Hans Rosling (whom I was able to meet in person), John Medina, Neil deGrasse Tyson, Tableau employees/customers, constant opportunities for networking, learning, and FUN…  Need I say more?  This conference ROCKS in a very unique way, from a former school teacher’s perspective–the Edu conferences I’ve attended are wonderful learning and networking opportunities, but certainly not as FUN as the Tableau Conference!

if you are not familiar with the Tableau Conference, take a look at the TC14 Agenda and you’ll get an idea.  There really is nothing like it.  The employees are passionate about their product, and we as customers are just as equally passionate about finding ways to utilize it creatively.  This makes for an excellent conference, with more material than one could ever hope to digest in one short week. Many of us will spend hours pouring through the material post-conference!

I Have a Lot to Learn

While I did not get to attend every single session I wanted to see (this is nearly impossible), I saw enough to open my eyes to the fact that my work in learning Tableau is nowhere near finished.  I’ve learned a quite a lot in the past 12 months about Table Calculations, and other “tricky” concepts, mostly thanks to folks like Jonathan Drummey and Joe Mako–who have worked diligently to help others understand undocumented and difficult-to-grasp concepts in Tableau.  However, I saw several sessions that focused on features that allow us to avoid the complexities and limitations of Table Calculations.  These included utilizing sets, data blends, and other Tableau concepts to accomplish what I may have “brute-forced” with a Table Calc in the past.  I’m particularly interested in the duplicate data source self-blend techniques demonstrated by Jonathan and others, like Bethony Lyons, a Tableau Product Consultant based in the UK.  I cannot wait to review the session videos and learn more (hmm, maybe a second blog post on this technique is in order?)

Being named a Zen Master can be intimidating, because so many of the Zens have pushed the boundaries of what can be accomplished with Tableau.  However, this group of users is humble and understands that Tableau is a tool that encompasses a HUGE number of features… and no one Zen is an expert in all areas of Tableau.  We all share attributes of the Zen requirements: Product Mastery, Innovation, and Teaching/Sharing — but we all do this in different ways.  This is important for anyone looking to Zen Masters for help in Tableau:  none of us know everything about the Product, but we all do our best to guide you toward your own understanding of the Product.  In addition, many of us are Zen Masters by night, supporting our mystical habits with 9-5 jobs! I hope to post more on becoming a Zen Master and how it impacted my TC14 experience in a subsequent post.

In Closing

I would have liked to have mentioned many more folks in this post–I met so many new people at the Conference, it would be impossible to thank them all here.  Just know, if you’re in the Tableau Community of users and employees, I appreciate all that you do to help others utilize this tool!

I may focus the next post on some of the highlights of the conference, as a new Zen Master: comments, thoughts, ideas? Share them in the comments below, or reach out via Twitter: @mluttonBI. This being my first post, I’m wide open to ideas for improvement, and would love to hear your thoughts on how to make the most of this little space on the web.

Until next time.