Best way to filter on trends in a chart

Dr_Duru
Dr_Duru Member
edited March 2023 in Datasets

I have a chart with a very large set of series values. I want to filter to focus on members of the series exhibiting specific trends (uptrends or downtrends) based on the y-axis value. Does Domo provide any such filters? Or perhaps someone can recommend a Beast Mode calculation I can reuse?

Best Answers

  • MarkSnodgrass
    Answer ✓

    Not sure if this is what you are going for, but I put together how to get the Top N of a Subcategory that may help you.

    https://youtu.be/hfIUQqeRVkM

    Hope this helps.

    **Check out my Domo Tips & Tricks Videos

    **Make sure to <3 any users posts that helped you.
    **Please mark as accepted the ones who solved your issue.
  • GrantSmith
    GrantSmith Coach
    Answer ✓

    You could utilize a simple linear trend slope calculation however you wouldn't be able to do it in a beast mode because you'd need to aggregate some aggregations.

    sum((x−avg(x))*(y−avg(y)) )
    
    /
    
    sum((x−avg(x))*(x−avg(x)))
    

    This will give you a larger number for faster increasing trend and smaller numbers for slower increasing trends. + values would mean the trend is going up, negative would be going down.

    You'd need to use a window function to calculate your average across your partition for each row then use a separate tile to calculate the rest.

    **Was this post helpful? Click Agree or Like below**
    **Did this solve your problem? Accept it as a solution!**

Answers

  • MarkSnodgrass
    Answer ✓

    Not sure if this is what you are going for, but I put together how to get the Top N of a Subcategory that may help you.

    https://youtu.be/hfIUQqeRVkM

    Hope this helps.

    **Check out my Domo Tips & Tricks Videos

    **Make sure to <3 any users posts that helped you.
    **Please mark as accepted the ones who solved your issue.
  • GrantSmith
    GrantSmith Coach
    Answer ✓

    You could utilize a simple linear trend slope calculation however you wouldn't be able to do it in a beast mode because you'd need to aggregate some aggregations.

    sum((x−avg(x))*(y−avg(y)) )
    
    /
    
    sum((x−avg(x))*(x−avg(x)))
    

    This will give you a larger number for faster increasing trend and smaller numbers for slower increasing trends. + values would mean the trend is going up, negative would be going down.

    You'd need to use a window function to calculate your average across your partition for each row then use a separate tile to calculate the rest.

    **Was this post helpful? Click Agree or Like below**
    **Did this solve your problem? Accept it as a solution!**