topic Re: Previous N week data in DAX Commands and Tips

Previous N week data

Anonymous — Fri, 05 Jun 2020 10:57:54 GMT

Hi,

I have a table that captures registrations for events and this table has 2 date fields - RegistrationDate and EventStartDate.

I need to calculate the cumulative No. of registrations for each event from the previous 52 weeks until the event starts (illustrated below).

ESD = Event Start Date

EventName	ESD-52week	ESD-51week	ESD-50week	...	ESD-2week	ESD-1week	ESD-0week
Event1	0	4	7	...	15	20	44
Event2	7	7	27	...	35	42	54
Event3	15	21	33	...	35	47	66

The last numbers (44, 54, 66) are the total registrations for the respective events.

Calculation on the registrations need to be on RegistrationDate and reaches its max at EventStartDate.

A tabular representation would work fine for me.

Any ideas on how this can be achieved, please?

Thanks in advance!

Re: Previous N week data

Anonymous — Fri, 05 Jun 2020 11:10:37 GMT

Hi Anonymous ,

https://community.powerbi.com/t5/Community-Blog/Week-Is-Not-So-Weak-WTD-Last-WTD-and-This-Week-vs-Last-Week/ba-p/1051123

Please check this blog.

Regards,

Harsh Nathani

Re: Previous N week data

Anonymous — Fri, 05 Jun 2020 11:22:31 GMT

Hi Anonymous ,

The criticality of my requirement is that we have 2 separate date fields and the calculation is based on both.

No. of registrations are being calculated on the RegistrationDate, whereas the offset needs to be calculated on the EventStartDate.

Re: Previous N week data

stevedep — Fri, 05 Jun 2020 20:08:25 GMT

So looking back 52 weeks but not before the registration date, or will there not be any registratios anyway before the registration date?

Re: Previous N week data

mahoneypat — Sat, 06 Jun 2020 01:09:39 GMT

The simplest way to do this would be just add a calculated column to your table with this formula

RegistrationWeek = DATEDIFF(Table[EventStartDate], Table[RegistrationDate], WEEK)

That will give you # of weeks before the event (-1,-42, etc.). You can then make a matrix with Event on the rows and the new column on the Columns, and use a running total measure like this

Cumulative Registrations =
VAR __thisweek =
SELECTEDVALUE ( Table[RegistrationWeek] )
RETURN
CALCULATE (
COUNTROWS ( Table ),
ALL ( Table[RegistrationWeek] ),
Table[RegistrationWeek] <= __thisweek
)

If this works for you, please mark it as the solution. Kudos are appreciated too. Please let me know if not.

Regards,

Pat

Re: Previous N week data

Anonymous — Sun, 07 Jun 2020 13:53:58 GMT

mahoneypat

This is excellent Pat... Thanks a ton.

Your solution is just about right. I can modify the rest as per the requirement.

Re: Previous N week data

Anonymous — Thu, 11 Jun 2020 13:19:18 GMT

Hi mahoneypat ,

As per your suggestion, this is what I could achieve.

There are a couple more things that I would like to know :

Question 1 - How can I populate the empty cells with the same previous value? (For eg - the highlighted part should contain 150 for -3, -2, -1 and 0 weeks)

Question 2 - If I select only events that has gaps, even the header (RegistrationWeek) skips the columns that has no data. How can I keep this static to always show 0 to -52 even if there is no data? (this is related to Question 1 as there is no data to show in that column)

Re: Previous N week data

mahoneypat — Thu, 11 Jun 2020 22:58:59 GMT

The reason you are seeing blanks is that there are no rows for that week, so there is not a RegistrationWeek value to go in the first part with SELECTEDVALUE(). To get around that there are two approaches that could work. I recommend the second one:

1. Add blank rows in the query editor to fill in the missing dates/weeks (e.g., make a full list of dates/weeks, convert to table, and then merge your data into it for each customer)

2. Make a disconnected table of regisration week numbers (e.g., with GENERATESERIES(-42,1,1) or something like that and use that in your columns. Then modify the measure expression to get the SELECTEDVALUE() of that column.

If this works for you, please mark it as the solution. Kudos are appreciated too. Please let me know if not.

Regards,

Pat

Re: Previous N week data

Anonymous — Fri, 12 Jun 2020 12:10:15 GMT

Went by the second approach and that helped... Thank you! 😊