Display new properties in the Radar plot appender node

KNIME Analytics Platform
1

Is there a way to display the new properties automatically in the Radar plot appender node. My set of properties to display changes and the new ones aren’t displayed by default. This is error prone since the node has to be configured each time the properties change. And even more problematic, it prevents me to use the node in batch. Did I miss something? Or the only way is to use D3 or Plotly library in the Generic JavaScript view node?
Best regards,

Jean-Christophe

2

Hi Jean-Christophe,

Thank you for reporting that issue. Would the KNIME node Radar Plot (Plotly) be an option? You can find in on the KNIME hub here: https://kni.me/n/T_F-9Hfsbh5Oqu-D. This way you can avoid the the Generic JavaScript node, but still have the functionality you mentioned.

Best regards,
Jeany

1 Like
3

Dear Jeany,

Thanks for your answer and the pointer. I have tested the Plotly node and it could be indeed a good alternative. It doesn’t have the same functionality to control the scales and property ranges though. I will run some tests to check if these limitations are blocking in my application. If I can’t use the Plotly node would you have a Generic JavaScript node example generating a radar plot that I could use as starting point?

Best regards,

Jean-Christophe

4

Hi Jean-Christophe,

Setting the scales and property ranges dynamically is a bit tricky as you already need to know something about your incoming data to do that. I had a chat about this with my colleague Ben and he had another suggestion that might allow you to keep using the Radar Plot Appender node. Do you know by any chance the maximum of properties (column names) that will be used in the plot at any time? You could configure the node with those columns, and then you should able to use the node with any smaller table (subset of the initial table) you want without reconfiguring the node.

Best,
Jeany

5

Dear Jeany,

Thanks for your efforts to provide me with a solution.
I have looked into the range/domain question and I will just use the property domains of the input properties calculated with the Edit numeric domain node.

// Do you know by any chance the maximum of properties (column names) that will be used in the plot at any time?

Yes, it could be capped to 10 properties maximum.

// You could configure the node with those columns, and then you should able to use the node with any smaller table (subset of the initial table) you want without reconfiguring the node.

I have already noticed I could reduce the number of properties from a set the node was already configured with. However in my scenario the property names will change and I can’t use a preconfigured node as these property names can change at any execution. I couldn’t find a workaround for this scenario.
Best regards,

Jean-Christophe

6

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.