Preparing a data model for Sankey Charts in Power BI

Рет қаралды 22,813

Күн бұрын

Пікірлер: 18

@mikebeccaria Жыл бұрын

This is one of those videos that I will definitely need to watch a couple times. I LOVE the high level of content here that continues to challenge me and think more deeply about functional programming strategies.

@davidirias3459 Жыл бұрын

SQLBI never fails !!! thank you Marco

@karansujata 3 ай бұрын

this is brilliant and exactly what I am looking to achieve in my upcoming dashboard. many thanks ❤

@philfletcher5341 3 ай бұрын

This is so good! Thank you for the in depth explanation of how to prepare the data! I'm working with tables in a SQL database, so struggling to convert the DAX, but getting close. Thanks!

@AgulloBernat Жыл бұрын

mindblowing! I need to try it myself to comprehend it completely

@Quynh_Phan 8 ай бұрын

How can you fix the color formatting even changing/adding layers with containing many attributes inside?

@JojoLo-o6q 4 ай бұрын

Hi, lovely video! I had a question about the filtering capabilities. Is there a way to filter for a specific node and visualize all the preceding nodes that lead up to it? (ie in your dataset, if I filter for just subscription, is it possible to see a chart that includes the full flow: web contact > start trial > subscription. Currently, it seems that filtering for one node only shows direct or immediate prerequisites, but is it possible to show the entire multi-level path? Thank you!

@SQLBI 4 ай бұрын

It is possible, but it's not immediate. Conceptually, you should: 1) Create a disconnected table that allows you to select the node to highlight (as a destination) 2) Retrieve the ContactId values that have that value in the destination (like 'Funnel Sankey'[To] ) 3) Filter these ContactId values in the measure that count/filter the result I think this could be a good use case for a calculation item in a calculation group, so you can apply it to any measure in the Sankey Chart. If you use a calculation group, you could just use one calculation item for each Event Type, so you apply it directly as a disconnected table. For example, this calculation item (for Cancel - you can create one for each Event Type) works for all the measures (you could use only one of the filters in Result variable if you do not use both types of reports). Cancel = VAR _SelectedContacts = CALCULATETABLE ( VALUES ( Events[ContactId] ), REMOVEFILTERS ( Events ), 'Event Types'[Event Type] = "Cancel" ) VAR Result = CALCULATE ( SELECTEDMEASURE(), KEEPFILTERS ( TREATAS ( _SelectedContacts, Contacts[ContactId] ) ), KEEPFILTERS ( TREATAS ( _SelectedContacts, 'Funnel Sankey'[ContactId] ) ) ) RETURN Result

@JojoLo-o6q 3 ай бұрын

@@SQLBI hi thank you for the reply! Can I ask how I might integrate the calculation measure with the filter? I tried adding it under ‘filters on this visual’ for the sankey diagram, but this requires me to “show items when the value is…” I’ve put not blank as a default, but I don’t see any filtering happening. Thank you!

@rinmai9225 6 ай бұрын

Tks Marco. So usefull

@MarceloNieto-cl4tp 7 ай бұрын

Is there a reason why my Events table would create rows for approx each iteration of eventID? For example, my data has 15 distinct batches (not counting 99) where a person can land. 1-5 are the first level (your web and phone contact), 6-10 is the second level, 11-12 the third level, and 12-15 the last level of the sankey that I want. The first unique ID in my data produces rows for 1-11, 15, and two blank rows for some reason even though they should only have rows for 2, 9, and 12.

@SQLBI 7 ай бұрын

Try to reproduce the issue using the sample file in the article and write a comment in the article comments, if you can repro using the sample file we can help there.