Discovery: Better repo health dashboard UI

We’re maintaining the repo heath data in csv file. The current format of data isn’t informative and can’t be used for any decision based analysis. Therefore we did some discovery to find a better tool which can provide better dashboard UI to this data and help in analysis

Tools we considered

  • Metabase,

  • Redash

  • Apache Superset

  • Datasett

  • Jupyter

Tool we selected

  • Metabase

We selected Metabase as it is an open source tool. It will be quite cost effective if we host it on our end. 

Comparison with Redash

Redash is quite close to Metabase but it is quite inclined towards the usage of SQL queries. Metabase provides single click filters and some commonly used filters with easy access where you don’t need much SQL. But This doesn’t mean that Metabase does not support SQL. Actually it does if you want to apply some complex filter on the data. You can also save your SQL snippets.

Comparison with Apache Superset

Metabase UI is quite user friendly. It has a soothing effect unlike Apache Superset which has quite old fashioned UI as compared to Metabase.

Comparison with Datasett

Datasett UI is also not very user friendly. One needs to know different CLI commands to set it up. So unlike Metabase it’s learning time is quite big. It also doesn’t have a graphic visualization feature of data.

Comparison with Jupytor notebook

The jupyter (pivottablejs package) UI also doesn’t seem very user friendly. And it has limited features as compared to Metabase. Like it doesn’t support SQL queries.