splink
splink copied to clipboard
[FEAT] cluster_studio_dashboard - Option to display clusters grouped by dataset
Is your proposal related to a problem?
Currently the only way to distinguish between datasets in linked clusters is to colour by source dataset. This helps to identify individual nodes but doesn't give an overall view of how the datasets link to one another overall.
Describe the solution you'd like
An optional change to the force-directed graph could enforce some separation between dataset subclusters.
Potentially other related visual changes, like colouring dedupe edges by source dataset (links between datasets in black or other neutral colour)
Additional context
Likely sufficient to close https://github.com/moj-analytical-services/data_linking/issues/409