How to exclude the Other bucket from Horizontal Bar visualization

We're using Kibana 6.1.4 (via bitergia.io) to visualize community contribution data to a project.

One of the visualizations we're trying to set up is a bar chart of the top 20 contributor names ordered by their number of contributions.

We've set up a Horizontal Bar visualization that does just that:

Bar-chart-data

The "Others" bucket is added automatically in the visualization and we have no control over it. The only workaround we've found to diminish the impact it has on the scale of the Y axis is to check "Set Axis Extents" in "Metrics & Axes" to manually set the "Max" value of the Y axis to be near the maximum value of the contribution count of the top contributor, thus truncating the "Others" bar. Using the "Exclude" field under "Buckets" to exclude Others did not work.

In summary, we'd like to remove the "Others" bucket from the chart and display only the 20 buckets with the user names. Is this possible?

Hey @dplanella, I'm unable to replicate what you're seeing. Would you mind checking both of your JSON inputs which I've highlighted below:

Thanks @Brandon_Kobel for the quick reply.

Both JSON input boxes are empty (Metrics > Y-Axis > Advanced and Buckets > X-Axis > Advanced).

I thought this might be related to the relatively old Kibana version we have. But seeing that you cannot reproduce it, and digging deeper, it looks like we're not running Kibana per se, but a fork of it => GitHub - chaoss/grimoirelab-kibiter: Soft fork of Kibana, for the benefit of GrimoireLab. So perhaps this issue only applies to Kibiter.

In any case, all the source data is public, and the dashboard itself as well, so here is a link if it makes debugging easier: https://gitlab.biterg.io/goto/0f3e693d78c6c6947c6780254011d311

The others bucket is being hard-coded by the fork here: https://github.com/chaoss/grimoirelab-kibiter/pull/51/files

1 Like

Argh, then I guess there is no way to discard or hide the "Others" bucket in this case.

Thanks again @Brandon_Kobel for going the extra mile of even checking the fork's code, really appreciated!

I've ended up finding a workaround:

Add a Split Series sub-bucket, with a Filters aggregation. Then set the filter to * . This effectively shows all results and bypasses Others . See it here in action.

The only downside is that the legend's label shows "*" (adding a label to the filter ends up adding two legends, one for each bucket, which is confusing). I wonder if there a way to remove legends alltogether?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.