Graph between user-selected Dates based on relative scores

I have a dataset created from a recursive flow that's updated daily.  

It has columns that include:

User-ID   Date  UserScoreOnThatDate


I can easily graph a set of users between dates selected by viewers in a Quick Filter.  But the viewers would to see only users with scores that are higher or with scores that are lower when you compare the scores on the selected start and end dates.  


Do you have any suggestions for how to go about implementing that?




  • Can you post a screenshot of what the result is? If the users have an individual score on each day you should see their score on each one of those days. I don't think I'm totally understanding what you would be missing so I'm hoping the screenshot will help me.

  • Here is a screenshot of the current graph.  

    The viewers would like to see *only* users with a score that's higher or lower on 17th April than it was on the 3rd April (where those dates are chosen by the viewer).


    At the moment, there are many  users who have unchanged scores and they are not the main area of interest.



  • You might add a Lag function to your dataflow to get a prior user score. Then calculate the difference between a row and its precedent. You could then add a filter to the card that displays only when that calculation is not equat to zero.



  • But if I add such a function to the dataflow, then would it not be a fixed value?  We'd like the viewer of the card to choose the range of dates within which to compare.


    Is it possible to use the valuesin a quick filter set by the viewer within a calculated field and then use that in a filter?

  • If I'm understanding you correctly, you could add both the current row score and the lag row score to the final dataset using the dataflow. Then create a BeastMode calculation to subtract one from another. You could then filter on the BeastMode value not equal to zero.

  • I think at the heart of my question is how can I let the user set the size of the "lag"?  


    E.g. If user-A has consecutive day scores from 1st April of 5,4,6,4,5, then the viewer interested in her if he sets the start date to 2nd April and the end date to 3rd April (or 5th April).  He is not interested in her if he sets the start date to 1st April and the end date to 5th April.


    So,if I understand your suggestion as "lag"  being equivalent to "the score N days ago", I dont see how that helps when a viewer of the card is setting the dates.


    I have the scores of each user for every date.  I just don't see how I can relate that information in the dataset to filtering on the card based on viewer-selected values.

This discussion has been closed.