Interactive Visualization / two interactive filters possible?

[Juergen]

Dear all,

I am struggeling with an interactive visualization in a component.
The workflow is attached.

The problem is:
The first interactive filter works, the second not… My expection is that the both filters are “combined” and that the respective graph reacts on both filter settings.

Does anybody have a hint for me to get that solved?

Thank you in advance
Jürgen
Test_Vis.knwf (17.2 KB)

I am using KNIME 4.0.0

[ana_ved]

Hi @Juergen !

I know what is happening: The way you are organizing the filters is correct, but violin plot uses the column cluster as grouping – and the information for column group filter is not reused. I am forward this as a request to our team.

Meanwhile, if you try using a different plot – like the parallel coordinate plot – should work.

Cheers,
Ana

[Juergen]

Thank you very much… I will try it tomorrow…

Interactive Visualizations using the free Knime version would be really great. I have done one exercise to implement a R Shiny app to make that happen, unfortunatly it consumes a lot of memory… but something like that would be highly appreciated.

Best regards,

Jürgen

[aabbaa]

hi @Juergen ,

Could you share how to use R Shiny in knime please.

Cheers,
Supachai

[Juergen]

Hi,

of course…

I have used an example from the Shiny Gallery, which you can find here:

Shiny - Plot plus three columns

As the example is an app with two files (ui.R and server.R), my example is “transformed” in a single file app (app.R). Shiny_R.txt (3.1 KB)
You can further develop this file within the Rstudio environment, as this will ease debugging.
The adaption for KNIME beside the single file app, ist given by these two lines:

dataset_all <- knime.in
colnames (dataset_all) <- make.names(colnames(dataset_all), unique = TRUE)

You have to copy the code in a R View Node. By executing the R View node a browser window will open and you can interactively play with the data.

Unfortunatly, you have to cancel the node execution after closing the Shiny app.

I hope this approach will help you.

Best regards,
Jürgen

PS: This setup is memory expensive. … and be patient, the browser does not open very fast…