Think about the users journey…

After attending a few Tableau User Groups and speaking to the community I always wanted to see what everyone’s ‘best practices’ were, what are the routes people take to get to the end viz? After attempting to complete a #WorkoutWednesday it became known that myself and Charlie go about some vizzes the same way in getting to the end result. It was rather spooky, relate-able, great minds think a like moment with how close both our blogs were. Therefore, here is my thought process…

When I am handed a requirements request I like to draw out ideas in my head or on the nearest scrap piece of paper I can find. This helps the requester understand what it will look like and give the full picture to work too for myself. It could range from 4 – 6 boxes built up of line and bar charts. It lets the requester say ‘No I’d like the line graph on the left, the bar at the top’ etc. It helps them as much as me; requirement gathers are by far one of the best part in this journey.

Here is some rough pictures I have used before…

When I get to the data side, I go through to understand what measure and dimensions I will have to play with and ask myself the following questions:

  • What makes up the column names?
  • What does half of these columns mean – I do not know everything that goes through my work place so it is okay to ask questions.
  • What date fields are we having to take in to consideration?
  • Do I need to add many calculations or are they already there built in the data source?
  • Do I need to rename any fields to make it easier?
  • How many rows are there?
  • What is the end goal?
  • What is it they want to see but cannot explain in words?

This starts to build up a picture of what I am about to work with – sometimes it is going down a rabbit hole and you have no idea but, if you think of questions to ask yourself you can be pretty prepared for it.

It is important to keep the data story you are creating efficient and to make sense to yourself as well as the end users. Keep the data points away from being over complicated, use tool-tips to the max to help understand, make the titles of the charts dynamic so when you have a filter or parameter selected they all change – like magic! It might seem small having dynamic titles but I always get great feedback on this and it helps remind what you are looking at. I am always told to think about the user’s journey. Not everyone can understand a cross tab full of figures and names, so we need to turn these multiple rows into a story so the users can easily understand and identity their data to make decisions on the fly.

Now, if I have to work backwards from a viz on #MakeoverMonday or a dashboard that I see and want to recreated, then I try to reverse engineer. As mentioned previously I carry my Tableau Bible everywhere with me and this is built up of pages full of calculations that I will not always remember. It is a great reference point; yes maybe a little old-fashioned hand writing them out.

If this ever got set on fire, I would be in trouble.

Reverse engineering now is my favourite way to learn new things in Tableau, and exploring different data sets. You get the moment of ‘who knew that could happen’. Learning from others is a great way to learn and pick up new tricks whether that’s blogs, videos, networking or trial and error.

Thanks for reading a small snip-it of how I get to my end Viz.

Am I winning or am I losing?

You will need a cup of tea for this, sit tight.

OK, so I had thought I would give the #WorkoutWednesday a go even though I feel completely out of my depth. I have already been watching Twitter seeing everyone’s tweets about the tool tips and LoD calcs so at least I am mentally prepared for the battle that is about to begin. Have I mentioned LoD calcs give me nightmares; I have been in dire need of using it but have not yet. I’m told I will need a use case for them, I understand the theory but not the practice – so let us hope today is a winner.


To start with, @Andre sent me the data in an Excel file; I plugged it in to Tableau to get a ton of years and was like what, this is not right. SO already, I went in and stole the data set from other workbooks so it is neatly laid out ready to go. I started with copying what was easily identifiable from the screen shot to my X and Y-axis (see above). You can tell it is a date field on the top, so I added the year and then used every measure to match the Y-axis. And no luck 😦

I left that and moved on to something else, in Andy’s instructions he mention you need to select a different stat – so that I knew was a parameter – dropped down his selection and wrote out the text in my parameter.


You already know with every parameter comes a calc so I knew that was my next step; an IF statement – I know some people have IF arguments, this worked for me and I knew it. #Winning number 1


After this I became a bit stuck, and unfortunately had to download the workbook and have a look, I didn’t want to completely undo myself I wanted to try and work it out from at least the measures before opening up the pills and seeing what made them. The first one was 25th Percentile – I was stumped, what does that even mean? Went to my trusty friend Google to get some answers and figured it was the 25% mark out of the 100% – but that was wrong too so I had to open the pill and look to re learn and figure out. Darn.

Next the pill Stat. So I knew this one – it is the calc we just made for our stats selection and that was showing the Median – if in doubt right click and hey presto it was there! #Winning number 2

It turned out as below, after I put the axis to dual – even my parameter was working #winning number 3.


So, what do I do now, why does Andy have his years on the top and long bars and a different number layout on the Y-Axis, what is this wizardry? Made a dual axis, sync’d it – again hey presto I had years on the top. #Winning number 4

I noticed there was a 75th Percentile – I mean I just did this one so surely it is the same calc but with 75? I tried it and it said our favourite message appearing ‘this calculation is valid’ yeessss #Winning number 5 I was pretty chuffed – did not know what it meant but it works, who cares. (Jokes I did care). I dropped that on the size marks card and we have lines, fat long chunky lines. #Winning number 6!

Then QBs in range – I saw this was on the tool tip card, I already knew people had trouble with the tool tips and I had heard there would be a LoD calc – holey moly I opened it and there it was in its glory. Let us just take a second, what? I get the idea yes I really do – but how to create that off my own back – no chance.


I played around with many of the features and tried to figure out some more of the calcs. The 75th TD was interesting, as I guessed it was using a different measure and noticed TD by luck – so duplicated all the calcs that needed TD (from AK’s viz) changed the measure to TD and they all worked. #Winning number 7 – I then had to rename a few of them measures to understand it more in my head, instead of saying TD this and TD that I edited the alias and would repeat the actions saying Touchdown this and Touchdown that… It works for me– if I see it how it is meant to do without shortcuts, I may remember it more.

Lessons learnt – what did I learn today? Well percentiles to start with, I mean so simple yet so complex. Then LoD calcs – let us not even go there, I think I have a fear! Thanks to @Andre at the end for explaining, I needed to turn my year into decimals – who knew? Although my X-Axis did not fancy the change clearly because it is still wrong.

Also to tick ‘none’ on the X-Axis for your tick marks to hide the years – sneaky. As well, I went over some calcs in the main viz and noticed a CASE Statement used – I used an IF Statement, why would they be different? Therefore, I took this moment to do some research to learn about the differences, concluded over developer preference…That could be wrong. Now I can use both – I am taking that as #winning number 8

I also learned I should be a little more confident in my work and findings because you do know, and you will figure it out and it is ok to cheat. It is better to cheat a little and learn new things then to cheat a lot and learn nothing.

So thanks Andy K – this one is for you.


P.S. This might look like a fail attempt but I really tried, and actually learnt a lot from failing. What doesn’t knock you down builds you up – come at me next week!!/




Create a website or blog at

Up ↑