BI Stories

Author: dchissick

  • Judging Iron Viz 2026

    Judging Iron Viz 2026

    The results are in for this year’s Iron Viz qualifiers, so I thought I could share some thoughts about the judging process, and of course about my involvement in it.

    https://www.tableau.com/blog/top-10-qualifiers-iron-viz-2026

    The judges for the Iron Viz qualifiers are about 30 volunteers from the current year’s cohort of Ambassadors and Visionaries, obviously not including those who are actually participating. This year I volunteered for the second time (previous was in 2024) – I can’t take part in any case because of the legal constraints that limit Iron Viz entrants to residents of only 32 countries in the world (more about that here), so judging is a good way to be involved.

    The vizzes are screened by the Tableau team to make sure that they comply to the minimal standards, such as the chosen theme (this year it was “Food and Drink”), and then each entry is assigned to 3 random judges. Many of the entrants publicize their work immediately after submitting, but we are asked to try and ignore the social media noise before judging, for obvious reasons. I was assigned 11 vizzes, but I had to decline one of them, as it was created by a colleague from my company and I had even given feedback on the content, so it was re-assigned to someone else.

    We had more than a month from receiving the anonymized vizzes until our deadline, so there was ample time. I invested at least half an hour on each – first investigating it as a viewer, then opening the workbook and seeing how it was created, including data sources, calculations, worksheet structure, and actions, and then assigning the scores and adding brief comments. When I had finished with all of them, before submitting all the scores into an online form, I also looked at the final ranking and evaluated if it reflected my general impression of the vizzes – because sometimes you can get immersed in the small details and scores, without seeing the big picture: which viz is the best?

    I was impressed by the amount and variety of data collected by most of the participants. Most of the vizzes were based on multiple datasets with a wealth of data about their subject, and I’m sure a lot of effort went into collecting this data, before even starting on the design.

    I have to note that most of the subjects did not interest me at all – I’m the total opposite of someone you would call a “foodie”, and I don’t drink almost anything except water. Some of the data, and even the food and drink names, could have been pure fiction, and I wouldn’t have known the difference.

    I am not going to divulge any recognizable information about the vizzes that I judged, of course, only my general impressions. Overall, of the 10 vizzes, I would have said that one was of beginner level, 7 were good to very good, and two were outstanding.

    We judge all the vizzes by awarding scores from 0 to 5 on the well-publicized criteria: Analysis, Design, and Storytelling. This is what the scores that I awarded looked like:

    Not much to be learned from this: most vizzes correlated well with their story, only the good ones had some complexity/nuance, and accessibility is not on everyone’s mind.

    Another way of looking at it:

    And the final results (anonymized Viz number), out of a maximum of 15 points:

    My thoughts about some of the judging criteria:

    Analysis

    • Viz topic aligns with the contest theme.
      • This one is obvious, and you don’t get a score for it – just True or you’re out.
    • Dataset and calculations appear to be accurate and clean.
      • This is where I check the data model: are there lots of data sources with disorganized names, or is there a nicely structured data model? And are the calculated fields well named and structured?
      • Errors cost points here – if an action or calculation performs incorrectly.
    • Analysis illustrates profound insights grasped from visualizations.
    • Analysis supports the story being told.
    • Analysis has been mostly produced within Tableau.
      • I understand this as meaning that if you’re calculating aggregates in SQL, or bringing in coordinates for a Sankey chart from Excel, you’re not producing analysis in Tableau.
    • Analysis highlights a broad range of Tableau capabilities.
      • Here is where I gave more points for using Tableau’s visual features: interesting chart types, dashboard actions, show/hide buttons, DZV, and even recursive parameter actions (my favorite, of course).

    Design

    • Accessibility is applied in one or more of the following ways: colorblind/low vision (contrast) friendly palette, limited use of images to convey text, font size 12pt or larger.
      • I deducted points mainly for tiny fonts (there were lots of those), and a little for color palettes – it’s easy for me to judge, because I am slightly color blind.
    • Visual elements add to the overall understanding of the visualization rather than distract.
      • Don’t use charts if they don’t convey any information.
    • Interactivity and layout are user-friendly, instructed/specified, and purposeful.
    • Charts are clearly presenting the data.
      • If I need to investigate a chart for more than a few seconds in order to understand it, you lose points.
    • Charts contain a title, summary, and/or caption.
    • All charts contribute to the story.

    Storytelling

    • A clear story is being told.
      • Don’t just display data – you need to tell a story. Actually I’m not very good at this, but I can recognize a good story when I see one.
    • Story flows through visualizations and guides consumer from question to insight.
    • Visualizations and animations support the story being told.
    • The story includes a unique idea or perspective.
    • The story being told has complexity/nuance that elevates the visualizations.
      • I looked for something special in the story: is it a bare analysis of the data, or are you investigating (or making me investigate) and discovering something?
    • Storytelling captures and maintains interest throughout the entire viz.
      • This is the easiest: was I captivated from the moment I first opened the viz, or did I lose interest at some point? The first impression was most important here.

    So what about the results?

    The viz that I ranked 3rd was in the top ten.

    The viz that I ranked 2nd was in the top three, and qualified its creator to battle on stage in Iron Viz 2026.

    And the viz that I ranked 1st ….  nothing. Obviously two other judges didn’t like it as much as I did, or maybe Tableau have other considerations that also affect the final rankings.

    Summary

    Judging Iron Viz is quite a bit of work, but it’s one way for us Ambassadors to give back to the community. It definitely improves your critical thinking, challenges you to give constructive feedback, and of course it’s fun. Even the discussion within the closed forums among the judges is interesting. I’ll definitely do it again, if I have the opportunity, and huge thanks to Katy Clarke from Tableau who led the process and coordinated everything.

  • Multi-Fact Dimension Switching

    Multi-Fact Dimension Switching

    My session at DataFam Europe included a section about some limitations of Tableau’s multi-fact model, especially regarding stuff you can do in an “old” model but not with multiple fact (base) tables.

    A recent bout of troubleshooting with a customer unearthed some new revelations, and a partial solution, for the “Dimension switching” problem that I had demonstrated.

    Let’s start with the problem. I have a multi-fact data source, as below. The two fact tables have five common dimensions (Date, Currency, Promotion, Sales Territory, and Product), Customer is related only to Internet Sales, and Employee and Reseller only to Reseller Sales.

    I‘ll start by creating a parameter that will enable me to switch dimensions in my worksheet:

    Now I create a calculated field based on this parameter, which can be placed in my chart or table, and I get an error:

    Not good. In an old model, with just one base table, you can switch any field using this technique. Now, even if I’m only intending to use this “Selected Dimension” field with a measure from one base table (and I haven’t placed it in the worksheet yet), Tableau is already giving me an error.

    This is where the customer’s story comes in. One of the developers tried this, and received an error as expected, but another developer, working with a different data source (also multi-fact), managed to create a similar calculated field and use it in a worksheet. So we started investigating, and after some trial and error we now understand when the technique works, and, more importantly, why.

    I already know (from a conversation with Thomas Nhan, the Product Manager) that Tableau calculations in multi-fact data sources start from the base table. We’ll keep this in mind during the following examples.

    Step 1

    We start by referencing only a couple of shared dimensions from different tables (note that Category and Sub-Category are from the same table):

    That’s an error, because Tableau can’t decide from which base table to start calculating.

    Step 2

    We add a dimension that’s related only to one base table:

    No error, because Tableau can decide, based on the [Reseller Name] field, that it has to use the FactResellerSales base table.

    That’s the key: give Tableau a clue regarding which base table to use, and it will use it. Let’s try something else:

    Step 3

    That’s not working, because [Reseller Name] and [Customer Name] are related to different base tables. And it reminds us that dimension switching will only work with measures from a single base table, which we’ll see in the final result.

    But even if all the dimensions you need are related to multiple base tables – add a dummy field from the table that you’re referring to, just as a pointer:

    Step 4

    This is OK. Category, Sub-Category, and Promotion are all related to both Internet Sales and Reseller Sales, but Sales Order Number, which will never be used because “—” is not a value in the parameter list, is in the Internet Sales fact table, so it is telling Tableau to start calculating from that base table.

    And this is where you have to look out for the actual results. By pointing Tableau at a specific base table, we are limiting the use of the dimension to that table only – even though the selected dimensions are related to two or more tables:

    In the screenshot I have selected the Sub-Category dimension, but in the data pane it is applying to Internet Sales Amount only, even though it is related to Reseller Sales as well – because I “pointed” Tableau at Internet Sales with my dummy row in the calculated field. Note that Reseller Sales Amount even has the icon indicating a broken link, and hovering over it displays a message:

    Conclusion

    Dimension switching in a multi-fact model works, just as it does in the old relationship model with one base table, as long as Tableau is told explicitly which base table to use. However, there is currently no improvement in the capabilities, so you can’t switch dimensions in a single worksheet and display measures from multiple fact tables. Maybe that will come in a future iteration. And now you know how to “force” Tableau to choose a base table for the dimensions when necessary.

  • Don’t Extract an Extract

    Don’t Extract an Extract

    Recently, not for the first time, I was contacted by a Tableau customer: “We have a published data source, but after we refreshed the extract, the data in the dashbords isn’t updating!”. Why? Caching? A wrong connection? No – they had extracted the extract. As I’ve seen this mistake quite often, I’ll elaborate a bit.

    When you’re using a published data source (on Tableau Server or Cloud), the mode – live or extract – is defined in the data source. Afterwards, when connecting to the DS from a workbook, the connection has to be live, because you’re querying the DS directly in both cases. If it’s live, the queries to the database are sent through the DS definition, and if it’s an extract the data is queried from the extract on the server.

    The mistake is in thinking that because the data is extracted, the DS has to be defined as an extract within the workbook as well. That will actually create another extract, at the workbook level, that simply pulls all the data from the published DS and saves it within the workbook. And then, if you refresh the original DS data, the workbook is still using its own extract (which hasn’t been refreshed), and the dashboard data isn’t updated.

    So this is the correct configuration:

    In Tableau Desktop:

    And on the server it should look like this:

    And next time you encounter a workbook that’s “stuck” with outdated data, just make sure that you haven’t extracted the extract by mistake!

  • DataFam Europe 2025

    DataFam Europe 2025

    Note – all the links to session recordings go to Saleforce+, you need to register (for free) in order to watch them.

    Preparations

    I was a speaker at DataFam Europe 2025, which was held in London on 2nd-3rd December, so from my point of view the event started a long time in advance.

    The news about DataFam Europe was only released in September, and almost immediately the applications to speak were opened, on a very short timeline – from 16th to 29th September. I’m sure it’s not a coincidence that Agentforce World Tour London was held on December 4th, and that probably forced the schedule. Having spoken in 2024 I was well acquainted with the process, and to hedge my bets I sent in three different applications, all of them new. This involves filling in a rather long form, with the important parts being the title, abstract, and “Why should this session be presented?”. I knew exactly what subjects I wanted, but I used some GenAI to generate a list of catchy titles for each based on the abstract, and then chose and modified the best one. I then used it to refine the abstract as well.

    My application was approved very quickly (I was informed on October 8th), so I had almost two months to prepare the presentation. Tableau always assign you a “Content Owner” who makes sure you’re on track, but my previous experience had been that they fully trusted us, the community speakers, regarding the technical content, and didn’t ask to review it at all – which had surprised me the first time. This year I was contacted by a couple of Solution Engineers from Tableau a couple of weeks before the event, and we set up an hour to go over my presentation. Their feedback was really helpful, and I made a few changes, even though it was on a tight schedule. I don’t know if this was implemented for every community speaker, but it’s exactly what I would have expected, so it was very welcome.

    The Conference

    The action started with a meetup (nominally “Data + Women”, but open to anyone) at the Information Lab offices on Monday evening. This was mainly a gathering of data people chatting and playing games, where I caught up with quite a few online friends, and made some new ones as well. Notably, all the Tableau senior management who were in town also made an appearance.

    The Conference entrance

    Tuesday was conference day, and everyone appeared at Tobacco Docks – which is a relatively large venue, but a bit chilly (we were warned in advance). On the second day I brought the Tableau scarf that we received in our welcome bag last year (and it helped), but there wasn’t any similar swag this time. There was enough space so it didn’t become too crowded, but if you were looking for someone you could find them quite easily.

    I was the only attendee from Israel except for the Salesforce employees, but I connected with them at their presentations and we’re already continuing the conversations back at home. Apart from that I met many friends who I’ve encountered as an Ambassador over the past three years, some of them at my lowly level and others very famous in Tableau circles (various Andy’s, etc.), and the nice thing is that everyone in the community is treated as an equal, as has been noted many times before.

    Unlike last year, there was a pop-up shop with some nice branded stuff, at semi-reasonable prices, and I indulged myself. There were the usual partner booths, snacks throughout the day, and in general a very welcoming and friendly atmosphere.

    My loot

    After the sessions (more about that later) we had the evening reception, which was basically just a lot of people hanging around and chatting. No problem with that, of course, but at some point I just quit for a quiet fish & chips and back to the “hotel” at my sisiter-in-law’s house. Just a note for Tableau community managers – the huge Jenga game is a safety concern, you need to clear a 2.5m radius around it from seating options, because one day someone will get hurt!

    Wednesday was a shorter day, so attendees could travel home in the evening. I was staying anyway, so I was present when they packed up the shop and started giving away the remaining stock for free – which, IMHO, is an insult to those who bought the same items earlier at full price. That doesn’t mean that I didn’t take anything…

    The Keynotes

    There were three keynote sessions, with room for everyone (500 people?) to attend. First was the Opening Keynote, with the usual presentation by Tableau management. As usual for these days, the main presentation was mostly about AI and Agents. Then there was a great talk by Matthew Miller about some real Tableau work, on the humidity level in the soil for his trees (!), and a customer / community panel who said all the scripted things that were expected, but didn’t enlighten anyone in the community.

    I’ve read a lot of criticism (mostly on LinkedIn) about the direction, and I agree. The keynote seemed to be aimed at corporate executives who are thinking about purchasing Tableau (or upgrading to the AI capabilities), while most of the attendees seemed to be community members. Maybe the mix is different at TC, where there are thousands of attendees, but here it was the wrong audience.

    One phrase that I liked, also by Matthew Miller: “Accurate, actionable, analytical answers, for anyone, anywhere, at any time” (I hope I got it correct).

    Then there was DataFam Slam – basically an entertainment show pitting five Tableau employees against five community (DataFam) members with short Tableau tips, and the audience voting for each tip. For me it was mostly fun, but I’m sure some of the tips were new for many of the audience, and I’ve written in my previous blog post about Andy Cotgreave’s winning tip for the DataFam (lucky that he switched sides earlier this year).

    DataFam won, of course 😊

    On day two we had True to the Core (not recorded?), with Tableau’s senior management fielding questions from the audience. Last year I was the first to ask a question, but this time I just listened. Some of the questions were really insightful, and I felt that the answers were sincere – especially when the question was “What question do you not want to be asked?”!

    I’m not good at remembering this stuff (questions and answers), but I think Tableau (Salesforce) are trying to balance between maintaining and growing their core business – Tableau Core, which includes the DataFam community – while also catering for the Salesforce imperative, which is to increase adoption of Tableau (Next) within the 70%+ of Salesforce customers who don’t have Tableau, and can contribute a lot of revenue. If we keep that in mind, we’ll hopefully be able to thrive within both worlds.

    My Session

    I presented my session (“Tableau Multi-Fact Models: Insights, Issues, and Fixes”) on Tuesday afternoon, to a relatively full hall. The onsite technical team were very professional, and there were no glitches. Because it was a “Silent disco” format, where the audience had headphones for the audio, there was no real option for any interaction, which was perfect: I knew from the rehearsals that I was going to be just over the allotted 30 minutes, and I ran through my content very fast.

    On stage

    My aim was to present an overview of the multi-fact data model, that was released in Tableau 2024.2 and is now starting to be more widely used. I reviewed the history of the Tableau data source models, showed how the new features work, gave some tips on implementation (Insights), and ran through some residual bugs (Issues and Fixes) that will hopefully be fixed in the near future.

    Feedback was good, both from attendees that I don’t know personally and from fellow ambassadors (and a Visionary or two). I’m glad that I succeeded in enlightening some people, and I hope this will drive adoption of this model – and maybe push Tableau to invest a bit more in improving it.

    The recording is here.

    The Breakout Sessions

    Most of the time at the conference was spent attending the “breakout sessions”, 30 or 45 minutes each, with a choice of 3-4 different sessions in each time slot (easy compared to Tableau Conference, where you may have to choose between 20 at the same time). There were longer hands-on training sessions as well, but I chose to skip them, so as not to take a spot from someone less experienced, who could benefit more. They are usually really fun, but less intensive, and I know most of the stuff.

    On the first day, not by design, all the sessions I attended were focused more on Tableau Next and AI, and mostly presented by Tableau employees. At some point it became boring – and they even used the same demos, so I was seeing the exact same functionality twice. Fortunately, much of the content was relevant for me, because I have to understand the whole Tableau ecosystem in order to educate our customers and point them in the right direction. The progress being made with MCP is especially interesting, because it could enable integration between Generative AI and Tableau, without investing in Tableau Next or paying a steep premium for Tableau+ licensing. At the end of the day I decided that tomorrow I would attend only Tableau Core sessions.

    One of the MCP slides

    So on the second day I did a round of “real” Tableau sessions. The highlights:

    Tremendous Tableau Tips by Heidi Kalbe, Nhung Le and Tore Levinsen. Just a selection of tips, some well known, some less, but all interesting.

    The Secret Life of Tables by Agata Mezynska. A great session on how to make tables more attractive, and insightful, for the users.

    Agata on Stage

    Beyond the Boundaries of Tableau by Tristan Guillevin. Not content with his “simple” viz extensions, Tristan is planning new and exciting implementations on Tableau APIs not yet released.

    Co-Designing with AI by Pablo Gomez. Very relevant for old-school Tableau developers who want to use AI as a design assistant, not directly within Tableau.

    Of course I couldn’t attend every session that I wanted, so I prioritized the more advanced stuff over the fun, and I’m looking through some other sessions on Salesforce+.

    Summary

    Was the trip worth it? Definitely – but I was there on the cheap compared to most others: Speakers don’t pay for registration, I wasn’t paying for a hotel, and my employer payed for the flight.

    But ignoring the cost (or lack thereof), it was a good conference. Great networking, enough sessions for both the desktop and the AI crowds, and a constant feeling of community. I’ve heard from veteran conferencers that the community and content were better in the previous decade, but I accept that times are changing, and there’s enough to go around for everyone. I’ll make my bid for TC 2026 (much more expensive), and maybe I’ll be there. If not, we’ve been promised that there will be a DataFam Europe this year (2026) as well.

  • Text Bars

    Text Bars

    The most memorable Tableau trick that I encountered at DataFam Europe this year was something that I already knew of (vaguely), but seeing it twice brought it to my attention and now it won’t be pushed to the back of my trick cupboard again.

    This was presented both by Andy Cotgreave in DataFam Slam (minute 36:00), and by Nhung Le in Tremendous Tableau Tips (minute 20:00), with different context (you need to register for Salesforce+ to see the videos). I’ll discuss just the technique, and what it can be used for.

    So what are text bars? You can use a text field in Tableau to create a horizontal bar, and also manipulate its length using the data. This is possible due to the existence of the Unicode characters “█” (U+2588: Full Block) and “▒” (U+2592: Medium Shade).

    Let’s say we want to display our count of orders per customer, which is not a very large number, in bars within the table header. We create a calculated field:

    REPLACE(SPACE([No. of Orders]), ” “, “█”)

    The SPACE function returns a number of spaces as defined in the parameter that you give it (up to a limit, of course – don’t try to exceed 100 or so), and then we replace them with the Full Block character. The result is a text bar, that can be placed as a header in the rows of out worksheet:

    The label at the end is a simple concatenation:

    REPLACE(SPACE([No. of Orders]), ” “, “█”) + ”   ” + STR([No. of Orders])

    Another option is to use the text bar to display a percentage value, though it has to be rounded. In this example I am rounding to the nearest 5%, and filling up the 100% with the partially shaded bar:

    REPLACE(SPACE(ROUND(MAX([Profit Ratio], 0) * 20)), ” “, “█”)
    +
    REPLACE(SPACE(ROUND(20 – MAX([Profit Ratio], 0) * 20)), ” “, “▒”)

    Note the MAX function, which I am using to eliminate negative values.

    The end result is below. No concatenated label this time, because formatting the percentage value to a string is complicated, so I simply added the Profit Ratio value as the next field. The Segment field is there to show that the bars are within the header.

    Andy showed how text bars can be used within tooltips (before Viz in Tooltip was released), but I’m sure the community will think up other implementations, apart from the simple spicing up of a table that I showed, and probably using other special characters as well.

  • Tableau Support

    Tableau Support

    Most Tableau developers have little or no interaction with Tableau’s technical support, but as a partner consultant dealing with lots of customers, I have opened dozens of cases over the years. Here is a short history of the experience, and some tips for newbies.

    Back in the day, before the Salesforce acquisition in 2019, Tableau Support was quite good. Anyone could open a case, even without a Tableau login, and once you traversed the initial questions trying to direct you to existing answers – or if you had the direct link – there was a comprehensive form to fill in. This included lots of technical details: type of problem, version, operating system, and the like, plus space for a short description, a long one, the urgency of the case, and files to upload.

    Responses were relatively quick, and they usually knew their stuff. Sometimes the first response was by email, and sometimes on the phone – I even had the local number through which the calls were routed saved on my mobile.

    After the acquisition the quality of support dropped significantly – both in my opinion and what I heard in the community. Some cases got no response at all, others were left hanging, and there was pessimism all around.

    During this period, support switched to Salesforce, and you could only open a case after logging in to Salesforce and connecting to your Tableau account through there. That made opening cases more difficult, because many small customers (and some large ones) don’t bother with all the configurations on their Customer Portal. Luckily we have a partner account, so in many cases I open a case through our account for a customer, especially if I have direct access to their Tableau environment (because the call will come in to me, of course).

    However, the support experience started improving. A lot. Today Salesforce Support is, in my opinion (again), significantly better than it ever was, and better than the old Tableau support. Some parts of the process are different, but the responses are faster, and definitely very professional. Some examples:

    • I opened a Severity 2 (urgent) case for a bug that was a show-stopper for my customer, and received a phone call within an hour. They already had a similar problem logged, and within a day we held a short session to confirm, and my report was added to the original. Unfortunately it hasn’t been fixed yet, but that’s a product issue, not support.
    • Recently we had a total outage on a customer’s cloud site, where no-one could view any dashboard. I opened a Severity 1 (critical) case, and received a phone call even before I had added the technical details. The support rep stayed with us for 3 remote sessions over the next 24 hours, until we found the root of the problem (which was caused by a the customer’s network security).
    • Another small but irritating bug that I opened with Severity 3 was easily reproduced, and fixed almost immediately, in the next patch: opened on 7 July, fixed in version 2025.2.1 (22 July).

    The Process

    So how does it work now?

    First, you need to be logged in to your Salesforce account and linked to your organiztion, and then you can open Salesforce Help (https://help.salesforce.com/s/cases). There you see a window with the “Ask Agentforce” option:

    The “Ask Agentforce” window

    Clicking opens a chat window, and I simply ask the agent to open a case, supplying as many details as possible. The agent asks some follow-up questions, and then creates a case.

    In the chat you can’t supply a long a detailed description, add lots of technical details such as version, or upload any files. So immediately after the case is opened, open the case page (it opens automatically, or you have a link in the automatic email you receive) and add a comment and any files. You can also respond to the email, and that appears as a comment as well.

    Responses are quick, by phone and email, and I have the current incoming number (from the UK) on my mobile. Obviously they don’t know everything, but there seems to be a good knowledge base – in a recent case they couldn’t find the answer, so I asked on the forums, and was assisted by Diego Martinez (Tableau Visionary) and his memory of a previous case. In such a scenario it’s important to ask Support to add the additional knowledge to their KB, so it will be easier for future customers, and they were responsive to my request to do so.

    Feedback

    I still have an issue with the feedback, after the case is closed. In the old system we received a detailed survey asking about various aspects of the support – response time, representative’s knowledge, general satisfaction, and more.

    Now the survey focuses mostly on the AI agent, which is just for opening the case, and not on the support itself. There are just five questions:

    How do you describe your overall satisfaction with Technical Support for this case?

    What can Technical Support do to make it easier to resolve your issue?
    Please select up to 3 areas that most impacted your experience.
    (this appears only if you have selected anything less than “Satisfied” on the first question)

    How would you rate your experience with Agent Astro (AI-powered Autonomous Support Agent)?

    What could have improved your experience with Agent Astro (AI-powered Autonomous Support Agent) before escalating to a case?

    What else should our support managers know about your technical support experience?
    (comment box)

    Part of the feedback survey

    So 40% (or more) of the survey is about submitting the case, which is maybe 5% of the process. There’s no real option for rating and providing detailed feedback to the technical support, and it could be improved. I always leave comments about the fact that they’re only asking about Agent Astro and not the support itself, and I hope someone is reading them.

    Summary

    Tableau Support (through Salesforce) are very good, and probably still improving. Don’t overload them with questions about functionality that can be answered in the forums (just moved to Salesforce Trailhead), but don’t be shy about opening a case for real issues. And if you encounter something that looks like a bug, start by trying to find it in the open issues site, and if it’s there click on the report button so it gets more traction. Only open a case if it’s not there.

    Report that you are affected by an existing issue

  • Tabbed Views

    Tabbed Views

    Every time we publish a workbook in Tableau, to Server, Cloud, or Public, we have the option to select the “Show sheets as tabs” checkbox. Obviously, this affects the display of the worksheet/dashboard tabs at the top of the view, but what else? And how has it changed over the years?

    Permissions

    A workbook with tabbed views counts as one entity for permissions. Any user who can access one view, can see all the others as well. They can also move between the views using the tabs, of course – that’s why the permissions can’t be separated: the display shows all the published tabs by default, so the user has access to them.

    If tabbed views are disabled, each view in the workbook has its own permissions. Users can’t move between them using tabs, and they have to exit back to the workbook level first – so permissions can be set separately, if required. I won’t go into detail about how to actually set permissions at the view level, the documentation is here.

    This enables the owner of the workbook, or administrator, to fine-tune the permissions if necessary, so some users don’t have access to certain dashboards, even in the same workbook – and in certain scenarios this is definitely a relevant use case.

    Dashboard Size

    The lesser known effect of setting tabbed views is the dashboard sizing. When tabbed views are on, all dashboards will be displayed as the same size in the browser, and if some have different sizes, the largest is used for all.

    What does this actually mean? If all dashboard sizes are set to “Automatic”, then there’s no problem. But suppose that you have one long-form dashboard in the workbook, with a fixed height of 3000 pixels that enables the user to scroll down. If tabbed views are on, all the dashboards will enlarge to 3000 pixels in height. And the same goes for the width, of course.

    If tabbed views are off, there are no such limitations, and every view can have a different size.

    Navigation

    One of the features that has changed over time is the navigation between views. In previous versions of Tableau, turning off tabbed views disabled the filter actions that navigate between views. I don’t recall in which version this changed, but now all the navigation options work even when tabbed views are off: filter actions, go to actions, and navigation buttons.

    This was actually a serious limitation in the past, because you nearly always had to compromise on something – the permissions, size, or navigation. Now that there are no limits on navigation, it’s much easier to decide which option to choose.

    One exception, of course – if tabbed views are off, you can’t navigate to a view for which you don’t have permissions. There are different results: a filter action simply won’t do anything, while a navigation button is grayed out and will display an “Access denied” message if clicked.

    Performance

    Even though there are still some resources on the internet saying that when you open a workbook with tabbed views, Tableau calculates all the dashboards in the workbook, that is totally incorrect.

    There is no difference in performance between tabbed and un-tabbed views. Tableau loads a view only when it is accessed by the user, and it always loads the workbook metadata when you access the first view – because it is stored as a single XML file. I have verified this using performance recordings, but it was always rather obvious: I’ve seen tabbed workbooks with 20-30 dashboards loading the first one in no time at all, so there’s no way all of them were calculated within that time. So don’t worry, if you have performance issues they won’t be caused by the tabs.

    Summary

    Tabbed ViewsNo Tabs
    PermissionsSame for all viewsCan be set separately
    Dashboard SizeBy largest viewIndependent for each view
    NavigationOKLimited only by permissions
    PerformanceOKOK

  • How to Un-nest Arrays using Tableau

    How to Un-nest Arrays using Tableau

    This is a revised copy of my original post in the Tableau Community blogs from 2023.

    Over the years I’ve encountered many examples of data that includes arrays of values within a single field – either in JSON format or as simple text, such as this:  [[4,0],[5,3],[7,3],[8,0],[4,0]]

    Usually the data can be un-nested using SQL or JSON functions into a set of separate rows, which can then be linked to the original data table as a separate table. However, a few years ago I was working on a system with no such option: the original data was in ElasticSearch, and our ODBC driver didn’t enable any un-nesting functions. And we needed to get the row data from the arrays.

    It turns out that Tableau can do this quite easily. Here is the solution:

    1.      Build a regular data source from your table, with the array field as normal text. In the example, it’s called [Shape Coordinates].

    2.      Create a table of numbers from 1 to N, with N being larger than the maximum size (number of items) in your array. The table can be in Excel, the database, or any other format. Add the table to the data source and join it to the data table using a relationship, using a full cartesian join – meaning that you create a relationship calculation with an identical value (I use 1) on each side.

    Note – the example here unnests a two-dimensional array. Each value has any number of pairs of coordinates, specifying a shape to be drawn on the page. The real data actually had pairs of real coordinates (latitude/longitude) specifying a GPS track.

    Now, for every row of data we have numbers from 1 to N in the corresponding table, which we will use to find the 1st, 2nd, 3rd, and so on… members in the array, by splitting it using the commas and brackets in the string. I would have loved to use the SPLIT function here, but it doesn’t accept a field or variable as a parameter, so we’ll use FINDNTH instead:

    3.      We find the start of each member: Start → FINDNTH([Shape Coordinates], “],[“, [Num]-1)

    Note that I’m looking for “],[“, because the comma alone is not enough – it will find the split between each pair as well.

    4.     The end of each member: End → FINDNTH([Shape Coordinates], “],[“, [Num])

    5.      The length of each member is basically End minus Start (the last member has no End, so I use a maximum value, longer than any string that I would expect:

    Length → IF [Start] > 0 THEN

       IF [End] > 0 THEN [End] – [Start]

       ELSE 20

       END

    END

    6.      Now the actual split: Split → MID([Shape Coordinates], [Start], [Length])

    7.      And clean up all the brackets and commas: Pair → REPLACE(REPLACE(REPLACE([Split], “],[“, “”), “]”, “”), “[“, “”)

    I’ve now unnested my array and retrieved an ordered list of values for each row of data. In this case the values are pairs of numbers, but they could be any type of data. If there are less than N members in the array, the Split and Pair fields return Null and can be filtered out. The data looks like this:

    8.      I split the pairs into X and Y coordinates:

    X → FLOAT(SPLIT([Pair], “,”, 1))

    Y → FLOAT(SPLIT([Pair], “,”, 2))

    9.      Now I can visualize my geometric shapes on a worksheet using polygons:

    Discussion

    This technique works, and caused no performance issues on a dataset that included tens of thousands of records, with each array having up to 50 pairs of latitude/longitude values.

    The “normal” solution would unnest the arrays using SQL, thereby creating a new data table with a large multiple of the original number of records, though with very few fields (ID, X and Y). If you are visualizing only a small number of records at a time, I would expect much better performance from my technique, which calculates the array data on the fly and only for the specific records you need. However, if you are filtering or aggregating by array data from thousands or millions of records in one view, a pre-calculated table would probably be much faster.

    The sample workbook is on Tableau Public:  https://public.tableau.com/app/profile/dan.chissick/viz/UnnestingArrays/UnnestingArrays

  • Nested Table Calculations

    Table calculations are one of the most used features of Tableau, but some of their capabilities are less well known, and I’d like to dwell on one of these: Nesting.

    Nesting table calculations occurs when one calculated field references another. That’s not a problem in itself, but what happens when we want each of these to behave differently? Compute the first using Table (across), and the second using Table (down)? I’ve found out that many developers don’t know how to do this, so here is a brief explanation.

    Let’s illustrate it using an example, based on Superstore data. I want to display sales by Product Sub-Category and period, and highlight the top 3 increases in sales for each period.

    I start by creating a simple table with sales by Sub-Category and Year:

    The first calculated field is the difference between each year and the previous year:

    Period Change:     ZN(SUM([Sales])) – ZN(LOOKUP(SUM([Sales]), -1))

    Note that this can easily be created using the Quick table calculation menu below, but then the next step would be impossible.

    Now I use this field in another table calculation:

    Period Change Rank:    RANK([Period Change]) <= 3

    This should return True for the top 3 changes per period (currently Year). I drag the field to Color, but after I set the colors that I want (green and white), I can’t get the correct results using the “Compute using” options.

    This is where the nested option is necessary, because the difference is calculated across the table (by rows), while the rank is calculated down (by columns). I open the “Edit table calculation” dialog box:

    Note that now there is a “Nested Calculations” dropdown, where I can choose the calculation option for each field, and now I can set “Table (across)” for Period Change and “Table (down)” for Period Change Rank. The result is now correct – the highlighted green cells are the top 3 increases for each year.

    To summarize – table calculations can be nested in two levels or more, and it is important to know that you have the option to set the “Compute using” option of each calculated field separately, enabling various types of analyses.

  • Published Data Sources and Calculated Fields

    Have you ever tried to edit a calculated field in Tableau, and seen only the “Edit copy” option? This appears when you’re connected to a published data source, and the field was published together with the data source – so it can only be edited through there.

    So should we publish calculated fields within the data source? What are the advantages and disadvantages? And how do we “move” a calculated field to our workbook when necessary?

    Note that I’ll use the abbreviation “DS” for “data source”, as it appears quite a lot in the text.

    I’ll start with the disadvantages:

    • As already noted, you can’t edit a calculated field that’s published in the data source. In my opinion this is a major issue – from my experience in both development and consulting with Tableau, the process includes a lot of iteration and trial and error, and the “Edit copy” barrier is a real nuisance.
    • Similarly, “Replace references” won’t work either. I’ve encountered several situations in which someone replaced a field, and was surprised that the results didn’t change. On checking, we found that the referencing fields were in the published DS, so they couldn’t be affected.
    • Performance.
      This is a tricky one. But we once analyzed performance on a data source with a lot (I think ~100) of published fields, before and after moving them to the workbook, and there was a significant improvement when the fields were with the dashboard – not in query execution but in compilation time, which can take several seconds. I don’t know the exact reason for this, but the results are consistent: you can save 20-30% of compilation time when the calculations are not in the DS.

    Advantages:

    • If a calculated field is in the data source, you can re-use it in different workbooks. It also prevents different developers from modifying “standard” calculations.
    • Materialization – pre-calculation of some fields in extracts – may improve performance slightly, but is relevant only for very simple calculations.

    Recommendations

    Note that these are my personal recommendations, and other opinions may be available.

    • In general, avoid embedding calculated fields in published data sources.
    • Notable exceptions:
      • Basic row-level calculations, such as:
        • [Price] * [Quantity]
        • [First Name] + “ “ + [Last Name]
        • DATEDIFF(“day”, [Start Date], [End Date])
        • IFNULL([any field], “N/A”)
      • Standardized business logic calculations, that you are confident will not change over time, such as:
        • Profit ratio:  SUM([Profit]) / SUM([Sales])
      • Groups, when they are used to clean up or organize a dimension.
      • Remember that there is no such thing as “pre-calculation” in Tableau beyond row level, so any fields using aggregations, level of detail functions, parameters, and of course table calculations, give you no benefit when published in the DS.

      So what happens in real life?

      In many cases, we create a data source and perform lots of testing or iterations, or even develop dashboards in the same workbook. But then we need to split the data source and publish it separately, while leaving the calculated fields with the dashboards. Here is a tested and proven process for doing this:

      Our starting point is a workbook with a local data source (extract), any number of calculated fields, and some dashboards (my example has just one).

        Step 1 – Copy the data source to a new workbook:

        • Create a new worksheet.
        • Drag any field from the data source to the worksheet.
        • Right-click on the worksheet tab → Copy
        • File → New from the top menu.
        • In the new workbook: right-click on “Sheet1” and Paste.

        Step 2 -Remove all calculations from the data source in the new workbook:

        • Click the filter icon at the top of the data pane, and select Calculation.
        • Select all of the fields now in the list (or batches of several each time, if there are too many), right-click and Delete. If any warning messages appear, you can ignore them.
        • Note that if you wish to retain any calculated fields in the data source, then they shouldn’t be deleted at this stage.
        • Delete any parameters or sets as well.

        Step 3 – Publish the data source (you may have to sign in to the server first)

        I always save the workbook in which I developed a data source, so I can easily make modifications and republish when needed. Therefore, I recommend not to check the “Update workbook to use the published data source” checkbox when publishing, and to save this workbook under a suitable name (“xxx DS.twb”) when finished.

        Step 4 – Connect to the published data source

        • Return to your original workbook.
        • Data → New data source → Tableau Server
        • Select the newly published DS.

        Now your original DS, with all the calculated fields, and the published DS, should appear together in the Data pane. One has the local extract icon (), and the other has the published DS icon.() Note that the published DS has no (or very few) calculated fields, compared to the local DS.

        Step 5 – Replace the data source

        • Right-click on the local DS, “Replace Data Source”, and make sure the local and published data sources are selected in the dialog box.
        • Click OK
        • You should see that all the calculated fields from the local DS are added to the published DS.
        • Close the local DS, and rename the published DS if necessary.

        Now you have a development workbook that is connected to a published DS, but with local calculated fields – that you can edit freely. You can verify that only the fields that you retained when publishing are “locked” for editing.

        Summary

        Published data sources are standard best practice, and you need to decide where to place your calculated fields, but it’s useful to know that there’s a relatively straightforward solution for moving those calculated fields from the data source to the workbook, at any stage of the development cycle.

        As a bonus, while writing this post I thought of a tiny new feature idea for Tableau: find a way to mark calculated fields from a published DS as “locked”, so the developer can identify them in the data pane. Please add some upvotes, and maybe we’ll see it implemented at some point in the future:

        https://ideas.salesforce.com/s/idea/a0BHp000016LlntMAC/differentiate-between-locked-and-editable-calculated-fields