Get the historical curve of the Top 1 brand this month

[Anonymous]

Hi there,

I am working on a social listening project, and one of the specific tasks is to display the Top 1 mentioned brand of THIS month and the historical curve of the brand. Probably like this

I have a date slicer working on them, the problem is that I do not know how to achieve this 2-step filtering:

1. Let date filter the Top brand

2. Let the brand filter the curve

The date cannot work directly on the curve or else only a dot will be displayed.

Please let me know if anyone has a brilliant idea to solve this!

[dearwatson · Accepted solution]

Hi nuci ,

A way to think about the first measure is a filtered table (lets call it top_brand) which is a summary of months, with a calculated column [top brand] which is the top brand for that month. 

I am assuming we start with something like this:

To get the top brand for each month we can calculate this with some tricks:

Count of Mentions = SUM('Brand Mentions'[Mentions])

Top Brand Mentions = MAXX(VALUES('Brand Mentions'[Brand]),[Count of Mentions])

Top Brand =
CALCULATE (
VALUES ( 'Brand Mentions'[Brand] ),
TOPN (
1,
FILTER (
VALUES ( 'Brand Mentions'[Brand] ),
[Count of Mentions]
= CALCULATE ( [Top Brand Mentions], VALUES ( 'Brand Mentions'[Brand] ) )
),
'Brand Mentions'[Brand]
)
)

this will give you this output:

then you can create a calculated table for the above output with this code:

Top Brand = ADDCOLUMNS(VALUES('Brand Mentions'[Month]),"Top Brand",[Top Brand])

Then you can model it back into the model with a cross filter as per below

then when you select a month in the "Top brand" table the top brand will be displayed and will go on to filter the brand mentions table for history.

its a crude example but hopefully it gives you an idea of how to approach it.. gets a bit tricky..

heres a link to the pbix:

brand.pbix

[Anonymous]

Hi @dearwatson  - This looks extremely helpful! Let me check this out first. Thank you so much expert 😉

[v-piga-msft]

Hi @Anonymous ,

Have you solved your problem?

If you have solved, please always accept the replies making sense as solution to your question so that people who may have the same question can get the solution directly.

If you still need help, please share some data sample with table format so that we could have a test on it.

Best  Regards,

Cherry