<< Back

My love/hate relationship with Tableau edge-cases

As the community grows and new users join, published work on how-tos and cookbooks become more and more important. But as a community gets bigger the ability to stand out as a leader becomes more difficult. One solution to this is to publish what I’m going to term ‘edge-case’ work, as in pushing the software and its use to places never seen or even intended. While this can inspire I worry that far too often it can push new users down. Setting the bar so high and unachievable it frightens those who wish to step out of their comfort zone. Here’s how I think we can all, as a community, avoid that from happening.

 

Why edge-cases are so important to the community

 

Readers can learn so much from them Tableau Public is an open book (assuming you choose to share your workbook) and so any interested party can dive right in and explore your workbooks – learning from your techniques and applying them to their own work. Many of us learn best like this: exploring, tinkering and hacking. By giving these edge-cases the creators give the community the chance to build on and improve each others work and learn along the way.

They show complex problems in a fun way  Often it’s hard for bloggers to share the work they are doing in their day job. Creating fun edge-cases gives them the chance to show interesting and new ideas, which are often complex and have taken a lot of development, in a fun or interesting way. It makes it fun for them to do, and makes it fun to read too.

They often provide work-arounds for problems that are hard to solve in Tableau Sometimes edge-cases explore difficult problems, challenges we all have doing what we want in Tableau. These challenges range from creating the design or interface we need or technically achieving a complex filter. Edge-cases give authors the chance to show their solutions.

There’s an immediate buzz from seeing something new presented in the software I love Users love to see the creativity of expert Tableau users – it inspires them. I have often been at events and heard from users who have said “I loved what you did with that XXXX that was crazy – where did you even get the idea?” or “did you see that XXXX built a YYYY in Tableau?” . It’s a fantastic conversation starter and brings new people into the community all the time.

They give feedback to Tableau Complex workarounds give feedback to Tableau. They show issues users are having that need a more elegant solution or feature in the product. Tableau is committed to making the product easy to use, if they see users going to extraordinary lengths to achieve something simple they will take steps to minimise it in future versions.

 

Courtesy of XCD

(Courtesy of XCD)

Why can edge-cases be detrimental to the community

 

They detract from what Tableau is great at Tableau is the market-leading self-serve data visualisation tool, it is not a way to build software applications, or games. Losing sight of what the tool is great at, and forgetting how it should be used, risks watering down the message the community has worked so hard to promote over time “Tableau helps you see and understand data”. Full Stop.

If there’s nothing to learn When the author doesn’t share what they learnt and only presents a finished solution – what is there for to learn?

They can make even experienced users feel inadequate I have spoken to several users who have been demoralised by complex posts, feeling that Tableau is a very complex subject and they will never master it in the same way that I have. We owe it to the community to continue present easy solutions and distil difficult subjects into simple terms.  Zen Masters, ambassadors and bloggers should have this at the forefront of everything they do.

 

Presenting edge-cases: Recommendations

With this in mind I want to lay out a series of recommendations as to how I think we, as a community, can all lay out edge-cases in a way that ensures everyone benefits.

Lay out the reason for your post Set your motivation for the edge-case, why did you decide to considerable time and effort on this particular challenge. What is it you want the community to take away from this post? What can they learn? How can they take your ideas and apply them to their work? If you’re just building something complex for the hell of it then explain why you chose Tableau – there are lots of tools out there. Often these complex solutions require more work to create in Tableau than in, say, D3 (because you’re stepping away from what Tableau is good at) and so explain why you didn’t use them. Think of your audience and what they want to see – this may not be what you want to show them.

Point out the alternatives It’s important that new users are given an informed choice. Are there simpler (but less effective) ways of achieving the same thing? Is another product better at it? Have other people blogged about similar ideas?

Present the takeaways and learnings for the reader If you are presenting methods and processes which readers can take away and apply then make them central to your post – anything else is incidental (but will undoubtedly get lots of attention). Providing a how-to for any key learnings and perhaps presenting them away from the complexity of your edge-case will help people learn.

Remind the reader what Tableau is good for Remember why we’re all here, it doesn’t harm to remind people what Tableau is great at.

 

Example

Olivier Catherin does a great job of demonstrating how to build a Decision Tree in Tableau, his post steps through an explanation of what he has achieved, how it works and how others can build it. It inspires us with what we can do with Tableau and also simplifies the idea into a series of steps. Presented differently this may have looked complex, but Olivier has presented his solution effectively with a thought for the new user to Tableau.

understanding+1

Conclusion

We are an amazing community, what we create and share every day is the envy of communities across the globe. The temptation to go bigger and better (and crazier) than others is huge, but that should not be seen as the definition of a leader within our community. Effective leaders lead by example: sharing, rewarding and helping and they should aim to make others feel comfortable.

New users join our community daily, old blog posts and tips and tricks get lost over time. There are plenty of simple tips and tricks in Tableau, let’s not forget those as we grow.

Chris Love

Nottingham, UK

12 thoughts on “My love/hate relationship with Tableau edge-cases

  1. Great post Chris. Thanks for discussing such and important topic. I would like to make a few points. 1.) I think the purpose sometimes is simply to see how far we can push Tableau. There may be better/easier tools to build a Cnake game, but doing that in Tableau is impressive and draws people in. My daughters still talk about that game and occasionally play. What functional purpose does it server? None related to visualizing data, but it doesn’t need to. These kinds of things shouldn’t deter other people using Tableau for it’s intended use. 2.) The community builds on ideas, even when not fully developed. Showing something in one form often leads to more and more. My original sankey posts had terrible instructions, but it led to you and Olivier improving on it with other techniques, then Chris DiMartini took the concept in completely different direction. 3.) When pushing limits on things, it’s often difficult to write directions at the most basic level. For example, when writing about Javascript API, unless it’s a beginner tutorial, there has to be some assumption that the reader knows some basic knowledge of HTML and Javascript. When getting to advanced topic there are many of examples of this, ex. Boran Beran’s blog. I guess what I’m saying is that I’m not sure there is a one-sized-fits-all for how these posts should be laid out. Also, you wrote, “When the author doesn’t share what they learnt and only presents a finished solution – what is there for to learn?”. As you pointed out earlier in the post, people can learn by dissecting work in the Public realm. “Tableau Public is an open book”, this is key to me. This is how many of us have learned the things we have. We need to encourage people to share workbooks and continue to foster that. I understand the need to blocking downloads, but those visualizations should be the exception and not the norm. Thanks again for the thought provoking post.

    1. Hi Jeffrey

      I really appreciate your comments. I completely agree with a lot of what you have raised above. A lot of what you raise above falls into my own thoughts – which is why I used the the love / hate analogy. To your points:

      1.) I think the purpose sometimes is simply to see how far we can push Tableau. There may be better/easier tools to build a Cnake game, but doing that in Tableau is impressive and draws people in. My daughters still talk about that game and occasionally play. What functional purpose does it server? None related to visualizing data, but it doesn’t need to. These kinds of things shouldn’t deter other people using Tableau for it’s intended use.

      I guess the last sentence nails it for me, these things “shouldn’t” deter people from blogging about Tableau and sharing their experiences but I believe they do. I have spoken to people who feel inadequate compared to Tableau masters like Tamas (myself included) and who don’t want to share their “simple” dashboards. Does this mean we should pushing boundaries? No, of course not. Does it mean we should be careful to be introspective in the community to ensure there is a broad balance of use-cases? Yes.

      2.) The community builds on ideas, even when not fully developed. Showing something in one form often leads to more and more. My original sankey posts had terrible instructions, but it led to you and Olivier improving on it with other techniques, then Chris DiMartini took the concept in completely different direction.

      100% agree, but your Sankey posts had some instructions, they showed the effort that had gone into them, the thinking behind it. They evolved over time. They allowed others to grow and share. I think your Sankey posts are a perfect example of where edge-cases are done perfectly for the benefit of the whole community.

      3.) When pushing limits on things, it’s often difficult to write directions at the most basic level. For example, when writing about Javascript API, unless it’s a beginner tutorial, there has to be some assumption that the reader knows some basic knowledge of HTML and Javascript. When getting to advanced topic there are many of examples of this, ex. Boran Beran’s blog.

      I think there’s a trap we can fall into a community of thinking that the value comes from cookbooks, how-tos or finished articles. On the contrary I believe the most use to the community is the motivation, the failures, the “I tried this and then went this direction”. So while I agree it is difficult to write instructions for some advanced use-cases then I believe there is still a lot of value in talking about problems, issues, remaining challenges / unsolved problems.

      Also, you wrote, “When the author doesn’t share what they learnt and only presents a finished solution – what is there for to learn?”. As you pointed out earlier in the post, people can learn by dissecting work in the Public realm. “Tableau Public is an open book”, this is key to me. This is how many of us have learned the things we have. We need to encourage people to share workbooks and continue to foster that. I understand the need to blocking downloads, but those visualizations should be the exception and not the norm.

      I completely agree, but to my point above let us not lose sight of the fact that there is as much learning in the process of making a visualisation, and the thought process, as there is in the finished visualisation.

      Thanks for taking the time to comment, this is turning into a fascinating discussion and I don’t, by any stretch, have all the answers. So great to have lots of feedback on this.

      Chris

  2. Excellent analysis here Chris. While it’s always cool to see the edge cases, when you step back, there is usually a better non-Tableau way to have accomplished it since they are using Tableau beyond it’s intended purpose. If the author can point to specific techniques that were used in the edge case that are applicable to normal uses, then it certainly could be valuable. But if the intention is just to build something crazy just to prove it can be done, then the long term impact may indeed be negative as it leads to further misuse. I would liken it to MS Excel – its a spreadsheet program that smart people (including many who found their way to Tableau) were able to stretch way beyond it’s means through clever manipulation and coding. IMO, the misuse/overuse of Excel is what created the environment that allowed Tableau to emerge in the mid-2000s in the first place. We were all trying to use that tool (Excel) to do so much, that we got fell off the incremental gains curve and were wasting too much time in the tool. I see many of these Tableau edge cases leading down a similar path.

    1. Thanks for your comments Jason, I’ve been guilty as anyone in building some crazy stuff, and I think it still has a place within the community but we need to be careful about the balance given to them vs other simpler use-cases. Your excel analogy is interesting and I don’t think we’ll ever get the same situation with Tableau, mainly due to the very clear direction of the product and team behind it.

  3. I really enjoyed this article Chris, and when I read Jeffrey’s comments, I found myself mentally finding merit in both perspectives. I supposed I’m an upper middle-class Tableau user, with a lot more experience than a great deal of people, but I’m no Jedi either. I recognize that I find myself mentally in the middle of the conversation because that’s how my experience has been. I’ve seen workbooks that made me drop my jaw with their coolness (the 3D Tesla car comes to mind), while simultaneously acknowledging that I could find no practical use for the “how-to” element. I’ve sat under the shadow of the Tableau Sith and learned about really amazing things that I would never (and should never?) be bold enough to attempt. And I can also truthfully attribute the majority of my Tableau training to the habit of dissecting and reverse engineering other people’s workbooks. I have at times failed in those attempts, wondering “how the heck did they…” So I still land in the middle, which I suppose is to echo Jeffrey – “I’m not sure there is a one-sized-fits-all”… I’ve needed, and enjoyed, posts all along the spectrum from basic table calcs to “don’t try this at home”. If my eyes start to cross, I just move on. What’s really great about this community is that if there’s something that’s way over my head, but I want to dive in because it solves a problem for me, I know I can reach out to the author and they’ll more than likely be happy to help me understand.

    1. Hi Cathy, thanks for commenting, I really value your input here. I guess I will just echo my points in the response to Jeffrey – I don’t want us to fall into the trap that one-size fits all, or that the only way is a cookbook / how-to write up. I think people gain much more from write-ups focusing on the failures and problems in creating a visualisation than they do the final visualisation.

      So as a community let’s get more willing to talk about the failures.

  4. Hey Chris,

    Great article, and one that is very timely for me. I’ve been having a lot of “self-doubt” about my contributions to the Tableau Community lately. For the last year and half of participating in the forums, I had only one goal, help every user figure out how to build their viz. It wasn’t until recently that I stopped and asked myself, “Am I actually helping this person?”. What good is it that I help someone with complex table calcs on poorly structured data, if the end goal is a big multi-layered conditionally formatted crosstab that should have been created in Excel (Or not at all)?

    Yes, I am helping push what is possible in Tableau (And I’ve learned far more than I have contributed), but am I actually helping, or just adding fuel to the fire?

    If someone asks, “How do I build this Exploding Pie Chart?”, is the best answer “Follow these steps”, OR “Here are some great articles on the downsides of Pie Charts, as well as, some useful alternatives to visualizing your data”?

    At the end of the day, we are Data Analysts/Artists/Etc, not Tableau Developers/Hackers. Right?

    Best regards,
    Rody

    1. Hi Rody – great response. I love your honesty here with regard your self-doubt. I’ve had the same at points in my career – and I’ve come to realise there isn’t a single best answer. Are you helping someone if you don’t give them the answer they need? A lot depends on their situation, their Tableau maturity and their willingness to accept an alternative.

      I think the point is we’re all different and we use the product in different ways, as a community we should reflect that and not hold one group of use-cases above any others because they seem more complex or “clever”.

  5. Hi Chris,

    Loved the post, and your recommendations for what to include when presenting edge-cases are spot on in my view.

    Firstly – as controversial as this statement might be in some circles 🙂 – every well rounded analyst should acknowledge that Tableau, whilst easily my favourite tool, is not the best solution to every problem; not even every possible dataviz task! Even Excel is better at some things (perhaps nothing viz-wise to be fair), let alone the likes of D3, the graphics parts of R and so on.

    I see this problem in the real world amongst my peers where, when first introduced to Tableau, they try to recreate a bunch of stuff previously done in Excel as data tables, find it difficult, and give up.

    Most of the time they should really have thought about using best practice data viz to improve, rather than recreate, what they have in Tableau – but sometimes a data table or a different tool can be legitimately the best way to presenting data, or even basically a statutory requirement (in accounting, etc.).

    Surely the skill is to learn first best data viz practices, but then also what Tableau is good for. I’ve seen some great posts/presentations on this topic in the community, but perhaps something that needs emphasis when introducing new users – in a way that doesn’t dissuade them from the amazing capabilities that Tableau has.

    A few thoughts came to mind, first when trying to think of “legitimate” reasons for showing off edge cases.

    – especially in the corporate environment, users might not have access to many other tools. If the powers that be have bought into Tableau (especially Server) then, if you want to develop and share work internally, that might be the only realistic way. Therefore if you want to present a node graph, a Sankey or whatever, you might feel inclined to push Tableau beyond what it naturally does just to have something to easily show.

    (And then, having achieved something so clever, they blog about their adventure doing that – but never about the 100 far more insightful line charts they happily produced with conventional means!)

    – it might be a way to show Tableau what new features are desired. The ideas forum is the official place of course, but if Tableau see many posts with complicated solutions to ostensibly simple problems then perhaps people feel that influences them; c.f. many posts with crazy table calculations before LOD expressions existed.

    – human psychology is probably such that people want to share their most flashy, interesting, successful and clever work! Consciously or otherwise, it might seem “embarrassing” to share failure, mundane or visually unappealing work. It probably gets shared, liked, favourited and so on far less, which – although we all like to think ourselves above such things – many people in the world clearly are not! Few’s “Show Me The Numbers” is one of my favourite books, but some of my nearest and dearest laugh at how “boring” the cover looks in comparison to e.g. McCandless’ work (not wishing to restart that debate!).

    That said, I agree wholeheartedly with your point. It would be great to see more typical use of Tableau as well as the super fun stuff; the journeys people went through and what they learned, what problems they still have.

    Whilst I hope I can be considered a competent Tableau user there are certainly a couple of quite basic seeming issues I’d love to know people’s approaches on I should document, and it would be fascinating for others to do so also. The recent Twitter revelation that one can drag labels is an example of something in retrospect solving a basic problem, but it seems like it wasn’t known about 100% even amongst advanced users.

    One consideration I was having was, if it’s a “problem” based post, I wonder if it tends to be put onto the Tableau forums rather than blogs. I guess the forums are less visible to the general population, but certainly a great way to get help quickly from probably a wider audience than most beginner blogs have, leading to different incentives. Knowing when to blog vs when to use the forums / ideas/ etc. is possibly another discussion – interested to know your thoughts.

    Thanks again for the post,

    Adam

Leave a Reply

Your email address will not be published. Required fields are marked *