Better histograms

Hi

I'm trying to produce some pretty basic exploratory histograms, which are a couple of lines with ggplot.

Neither the native Knime node nor the JFree node does what I need. Are there any other capabilities/addons within Knime?

Example requirements (Titanic dataset): 

Age: continuous - to be binned

Survived (binary) - to be counted for each value (true/false)

1. Variation of survival with age.

--> Show as stacked bars or side-by-side bars

The nearest I can get for this is to toggle between two views, count and sum, which show total in age range vs survived in age range.

2. Variation of survival with age by sex (M/F)

--> Failed to do this at all.

Any help much appreciated - otherwise it will be back to Python (much as I prefer Knime).

Chris.

 

 

Hi ccfpcl,

There is no direct way to do this with the view nodes available in KNIME right now. However, it is possible using a combination of GroupBy nodes and the JavaScript Bar Chart.

For example, after binning the age variable, you could use the age bins in the GroupBy as groups and the mean of survived as manual aggregation. This can then be plotted in the JavaScript bar Chart.

Of course, if you want more options, you can always use an R Snippet node in KNIME to create the views via ggplot.

Cheers,

Roland

Hi Roland

Many thanks for taking the time to respond to me. I'll certainly give your suggestions a go.

Best wishes, Chris.