https://www.gooddata.com/ logo
Join Slack
Powered by
I may be missing something obvious: Can you have m...
# gd-beginners
e
I may be missing something obvious: Can you have metric filters on dashboards? They appear to be only possible within visualizations. Details in thread
As in, I have a dashboard with multiple visualizations. I want to hide all rows on the dashboard where 
spend = 0
. In each visualization, I can do this, but I want to show this filter to the user (and have it be modifiable at the dashboard level by the user, like other filters).
Here's how you do it in a vis, and here you see a special UI element
but you do not see it in these docs, or in the UI on the dashboard (only time or attribute filters)
To be clear, I have this metric filter in in each visualization of my dashboard and it works correctly. I simply want to show that filter on the dashboard for clarity and so it can be turned off as needed by the user
t
Hi Erik, metric filters are unfortunately not available on dashboards at the moment. Let me send your feedback to the product team.
👍 1
🙌 1
p
🎉 New note created.
🎉 New note created.
e
Sanity confirmed, thank you!
m
Hi Erik, to dig a bit deeper into this - would you use this feature only for excluding zeros or would you consider also other cases for dashboard metric filters? I believe a slight complication is, that the metric filter works on the granularity (combination of attributes) of each visualization - it is applied after the aggregation (similar to HAVING in SQL). And since on the dashboard can be visualizations with different granularities, the “same” filter might apply quite different filtering logic to each visualziation. For example if we had a metric 
No.Orders
in visualization 
No.Orders by State
. This filtered by 
No.Orders>1000
would return only those orders that are from states with over 1000 orders. (no matter what was ordered). The same metric in different visualization - 
No.Orders by Product Category
filtered by seemingly “the same filter” 
No.Orders>1000
would return orders of products from Categories that were ordered at least 1000 times. (no matter from what 
State
). (Both these screenshots are from the same data). These are very different groups of source rows (also the totals are very different). Would that be your expected behavior? And would your users understand what is happening with their data? Thank you for your insight.
e
Thank you for putting time on this, that is a great example. In my opinion, The feature is still valuable and valid, but would need to be considered carefully by the dashboard creator. As in, documentation and maybe a tooltip warning about this behavior would be needed. The metric filter could also be applied only to certain visualizations using the dashboard configuration if needed just like other dashboard filters. Above all, I lean towards clarity in my dashboards. Metric filters on the dashboard would have extra complexity/consideration in use, but that is mental effort put in by the dashboard creator so that the dashboard consumer can clearly and immediately see what the data they see means. Currently, data consumers will assume what they see is all the data when instead a "hidden" filter is being applied without any indication. Also, if the metric filter is visible, it will be editable. Right now, I have to have two versions of the same report copy-pasted for active vs inactive campaigns (spend > 0 vs spend = 0) because that filter must be hard coded into each visualization instead of being user editable
🙏 1
4 Views
Open in Slack
PreviousNext
Powered by