Multiple variables in clustering model charts for comparison
I would like to add a boxplot charts with multiple Legends & multi variables to MLAuto Cluster. However, IT seems the DSS chart function does not support it (Only Single Legend - single variable )
Any way to build a custom chart (e.g. by a python script) for clustering model comparison ?
Operating system used: Windows 10
Best Answer
-
Emma Dataiker, Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Dataiku DSS Adv Designer, Registered Posts: 52 Dataiker
Hey @Filip_Pejsa
,If the built-in charts are not sufficient for your needs you can always create custom charts using python!
The first step is to deploy your AutoML model onto the Flow.
Once deployed on the Flow you can interact directly with the saved model object OR apply it to a dataset so that you build your custom python charts off the results.
You can even save those custom charts as "Insights" that can be reused across Dashboards, etc. More info here: https://doc.dataiku.com/dss/latest/python-api/static_insights.html
Hope that helps,
Emma