0 votes


At my company we are currently evaluating different tools for our BI team and have to run some basic tests to offer them as possibilities. One of the questions we are trying to ask, we have been unable to solve with dataiku, and therefor wonder if we can have a helping hand in understanding the tool. 

So we are trying to achieve 'for the last calendar month the average count of users' logins daily.' Basically meaning if we got 3 users who have logged in the 1st of Nov say 3, 5, 7 times then the average for that day is 5.

Below is a snippet of my sortet dataset, but just to quickly go over it the columns needed are
appid (filtered to only show logins)
userid (users hashed value, always the same for the same user)
datetime_month (filtered to the last calendar month, November)

I have been able to produce a chart showing the total amount of logins for the different days in the November, but haven't been able to divide that with the number of unique userid entries for those days.

Any help would be very appreciated.

(ignore the coloring, no clue how to remove it - but that isn't necessary at this point)


1 Answer

+1 vote
Best answer
The grouping recipe or the SQL recipe with a group by function fit well your use case.
selected by
I looked into this and it worked, but not only that it also made a lot of loose rocks fall into place - so thank you :)
1,319 questions
1,339 answers
11,888 users

©Dataiku 2012-2018 - Privacy Policy