-
Notifications
You must be signed in to change notification settings - Fork 13.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Show continuous timeline on x-axis for Time Series plots #3708
Comments
Is there any plan for this? If inserting zeros for missing time series is acceptable, I can submit a PR for this. |
I think, this behavior is all the same with other time series representations. Effectively, it gets plotted what the DB query returns. If there is no data present for a certain interval step, then the DB does not return any row. I guess, the way superset is designed, you are supposed to resample the data via the Advanced Analytics tab in Explore in order to fill missing values. Personally, I would not favor to have the timeline generally filled with zeros, as often enough I am only interested in the real events. |
I agree, though, that this is counter intuitive to fresh users... |
Could it be an option to fill in missing data with zeros? Like a tick box? |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
Can we please re-open this issue? If superset didn't specialize in time-series data, I could maybe see that it might not need to fill in missing time periods with 0 values. This would just be plotting the data as-is with time being like any other category. However, superset does specialize in time-series data. And so it should handle the effective plotting of missing data for time frames. In time series analysis you need to represent all time intervals that you are plotting across. This is industry standard for time focused plots and it's what data viz viewers are expecting. |
Can we please re-open this issue +1. It's very easy to ignore the missing 0 value if the timeline is not continuous. I think we can add a function like 'padding timeline' in "Advanced Analytics". If user enable the function , a continuous pandas datetime series will be left join the queried dataframe and fill na with 0. |
Superset version
0.20.4
Expected results
When I display my data on time series plot, I want to see the continuous timeline with objects (bars or points) on it.
If there are any gaps between the objects, I want to see all the gaps on the timeline.
Actual results
Look at the timestamps below the plot for the Time Series Bat Chart.
All the objects are arranged side-by-side, so I don't see the gaps between them.
The same is for Time Series Line Chart.
The text was updated successfully, but these errors were encountered: