When you work with large data set and you want to be able to show everything in one viz at the lowest level of detail, then you probably want to have a dense view that at the same time will be easy to read. Here is a step-by-step guide for one chart that easily does the job.
First you need two hierarchical dimensions. In my case that is State and Postcode. The one higher in the hierarchy goes on the Columns shelf, while the other one on the Details mark. In this way you will create a view where you will be able to see each postcode represented by a small square grouped by the state it belongs to.
Next, you’d like to sort the dimension on your view by a measure of your choice. Here, I am sorting the Postcodes by the Socio-Economic Index, restarting the sorting for each State. Grab the measure and drag it to detail. Right click on it and choose the option Quick Table Calculation -> Rank. Then once again right click on it and this time choose Edit Table Calculation. I am using the Unique Ranking option and in order to restart the ranking for each state I make sure to select Compute using Specific Dimensions, tick the postcode and un-tick the State Name. Keeping the Show calculation assistance turned on, you’ll be able to see how this configurations reflect on your view (marked in yellow). For more information make sure to go through the original Tableau materials: https://onlinehelp.tableau.com/current/pro/desktop/en-us/calculations_tablecalculations.html#Basics
You probably noticed that nothing much changed. That’s ok! Even though it is not visible at first sight, every postcode got a unique rank. Check this through exploring the Tooltip.
Next I’d like to sort the postcodes by the rank. Right click on the postcode pill (the dimension you have on detail), choose Sort, then Descending using a Specific field – and chose the one we just used for the ranking.
Now let’s add the ranking calculation to the measures area because we’ll need it later. Drag the Calculation which you used for ranking from the marks card to the data pane on the left, under measures and give it a name. If you try to edit it, you will see how Tableau calculated the ranking based on the measure we provided.
Next, we’ll give the viz some color. Create a new calculation where you will tell Tableau to use different color for each set of 100 ranks.
If instead of this vertical look we want to create a horizontal one (which makes a lot more sense), we need to plot the squares (postcodes) on x and y axis that we have to specifically define. These are the calculations for the axis:
- X-axis: [Ranking]%10
- Y-axis: ROUND(INT(([Ranking])/10))
There’s one issue with the current viz. The first rank is not plotted correctly which distorts the whole picture. To fix this update the ranking calculation to:
RANK_UNIQUE(SUM([2016 Irsad]))-1
In order to make this view easier to read it would be best to add some white space and further emphasize the groupings of 100. We need to update the y-axis calculation.
This tells Tableau to plot each 10 rows of postcodes one row above where they previously were, starting from the second grouping. This leaves one empty row in-between.
With a bit more formatting you should be able to get to a view that will plot your disaggregated data in a way that will make it easy for the user to get some insights. The key here is to make good usage of the Tooltips providing detailed information about the data you present.
