FEATURE REQUEST: make updateable plots programmatically with python node in dataiku
tjh
Registered Posts: 20 ✭✭✭✭
We have a "streaming" or "batch data" use case with about 100 features. We would like to make 50 plots with 2 selected columns each. This is very tedious to do manually with the current only possibility in dataiku.
A notebook could be used to make the plots programmatically, however they would not update when the data is updated in the flow.
The only option that we see now is to use plotly within a python node.
So for the next release please provide an option to accomplish the above job.
A notebook could be used to make the plots programmatically, however they would not update when the data is updated in the flow.
The only option that we see now is to use plotly within a python node.
So for the next release please provide an option to accomplish the above job.
Tagged:
Answers
-
If you have Automation features activated, you can define Scenario steps to re-run a Python notebook and refresh it on a Dashboard. That would use plotly, ggplot, matplotlib or any of your favorite plotting library. Documentation is here: https://doc.dataiku.com/dss/latest/scenarios/steps.html#refresh-notebook-insight
-
We have added a feature in DSS 4.1 that may interest you: https://doc.dataiku.com/dss/latest/api/python/static_insights.html
You can now have programmatic charts through Python and update a dashboard automatically through a Python recipe. It also works with R if you prefer that language!
