7/12/2023 0 Comments Union tableau prepEvery time you adapt anything, Tableau will recalculate your view. You will be adapting a whole number of table calculations. I’m merely putting his screenshots into text, which allows you to adapt multiple table calculations at once without scrolling through a blog post in between.ĭrag Sankey Polygons onto Rows. I’m not adding anything new to Ian’s instructions in this step, so if you prefer to see screenshots, go to his blog post. You should now see a horizontal axis and no marks. Both should be computed along Path Frame (bin). Drag T onto Columns and Path Index to path. Make sure to have the three pills in this order, from top to bottom: Dimension 1, Dimension 2, Path Frame (bin).Ĭhange the mark type to polygons. And this piece of magic right here is what saves us from needing ETL and / or multiplying our data.ĭrag Dimension 1 onto Colour, Dimension 2 onto Detail, then move Path Frame (bin) to Detail as well. Now Tableau fills in the 96 points in between the two we calculated. How? Simply drag Path Frame (bin) into Rows, right-click, and select “show missing values”. We already created two of these 98 points (see Step 2). The idea behind this polygon approach is that we connect 98 points in a way that it looks mostly curvy. These two calculations are taken straight from Ian’s blog post. Set up an INDEX() calculation that computes along Path Frame (bins) as shown in the screenshot. Now create bins from Path Frame with a size of 1. I will call them Dimension 1 and Dimension 2, same as Ian did, as it will make the following steps easier. If your calculation doesn’t show up there, make sure to change its data type from date & time to date, then try again. There you can select under “when workbook opens” which date field to choose the allowable values from. Set data type to date and allowable values to list. I will call them Month Source for the start date, and Month Target for the end date. Set up two parameters for your start and end dates. If you don’t want to allow all values in there (e.g., only the last 2 years), adapt your calculation accordingly, so that your calculated field only contains the values you want to have end up in the filter. In case of the latter, if you want to allow dates at a higher level of date detail (meaning: less detailed), create a calculated field that contains a DATETRUNC calculation for the relevant date field and your chosen level of detail (e.g., week, month, quarter, etc.). Step 1: The basicsĭecide whether you want to allow all dates or a range of dates taken from your data. You want to be able to dynamically choose both points in time, and the Sankey should adapt accordingly. More specifically, you want to find out how the distribution of your customers throughout the different journey stages changed from one point in time to another, counting your customers per journey stage. You want to track your customers’ journey. I will be structuring my blog post using the same steps as Ian did, so you can always refer back to his post and find the same aspect in the instructions there. All this brilliance stems from Ian’s brain. I will freely admit that I don’t understand half the magic happening in the gazillion of table calcs we will encounter here. I publish these adaptions with his permission and in endless gratitude for his blog post. This blog post would not exist without Ian Baldwin’s fantastic instructions for a data prep free Sankey chart. If the latter is the case, do read on.Īre you fine with adding twenty new calculations to your workbook? Great, because I don’t believe there’s a way around that.Īll these questions cleared up, let’s delve into things. So either you are happy to take out one of those numerous templates mentioned above and reshape your data to fit that one chart type – or you are willing to make some concessions. As far as I am aware, that will need some data prep. If you answered No to both questions, please check out Ian Baldwin’s blog post because it has all the answers you need.ĭo you want your different statuses beautifully spaced from one another? Well, tough luck, because I cannot help you with that. If you answered Yes to these questions, read on. Is this the right article for you?ĭo you want to create a Sankey chart in Tableau? Do you want zero data prep for that chart? If you answered Yes to both these questions, read on.ĭo you need your start and end dimensions to be different points in time (different rows within the same column)? Do you need a non-summable aggregated measure (such as count distinct)? Read on to find out how to build your Sankey chart without any data prep. You want to build a Sankey chart in Tableau and have read through multiple how-to’s and looked at numerous templates, but they all require you to multiply and / or completely reshape your data, and you really do not want to do that for just that one chart? I’ve got you covered.
0 Comments
Leave a Reply. |