In order to analyze community structure we rely on Onion model. All panels related to community structure are based on the onion study included in grimoirelab-elk project.
This panel splits data into organizations. Onion model is computed individually for each organization.
In addition to Kibana filters and search box on top, filtering by
Source is allowed by using the top left corner widget. Notice the blue filter
on top used for showing only
Git data. This filter can be modified by using
Data Source widget mentioned above.
In Kibana/Kibiter, filters are combined by means of
E.g. if we enable Git and GitHub filters, under the hood we get
an ElasticSearch query like:
data_source=git AND data_source=github
Thus, by selecting more than one data source filter at the same time, we won’t get any result, because data come from one data source or another, but not several at the same time.
Data Source widget automatically manages this for you, so you just need
to select the desired data source in the drop down list and click on
Apply changes. Data is meaningful only if one and only one data source
is selected for filtering.
In order to have meaningful data, we need to select the organization we
are interested in from the donut chart on top or from
organizations table below bar charts. Then, a new filter will appear on top, next
to the blue one named
data_source:"git", and we’ll get the data we need in
LIMITATION: when no organization is selected for filtering bar chart on the top right corner will show group sizes —core, regular, casual— aggregated for all organizations. That is, if we have organization A with 5 people in core group for 2018Q1 and organization B with 2 people in core group for 2018Q1, bar chart will show 7 as value for core group, which is not the result of computing onion for both organizations together but the sum of their specific results. This is due to the use of pre-computed values by organizations. In order to get data for all organizations together, see Overall Community Structure panel.
Edit this doc