KNIME Shape Node Histogram Processing

knime-plotting-snippet.R

library(ggplot2)
library(reshape2)
library(plyr)
library(dplyr)

cd.df<- knime.in %>% 
	select(Label,starts_with("CentroidDistance")) %>%
	melt(id.vars="Label") %>% 
	ddply(.(Label),
	function(in.df) mutate(in.df,
		norm.value=value/sum(value),
		norm.vis=max(norm.value)-min(norm.value),
		std.vis=(max(value)-min(value))/mean(value)
		) 
	) %>%
	mutate(vcnt=as.numeric(variable),lcnt=as.numeric(Label)) %>%
	subset(!is.na(norm.value))

knime.in.clean<-knime.in %>% select(-contains("CentroidDistance"))
names(knime.in.clean)<-names(knime.in.clean) %>% llply(function(cname) gsub("\\s",".",cname))

comb.df<-knime.in.clean %>% merge(cd.df)
comb.df %>% summary

ggplot(comb.df %>% mutate(volgrp=cut_number(std.vis,4)),
	aes(x=vcnt,weight=norm.value,color=lcnt))+
	geom_density(aes(group=Label))+
	facet_wrap(~volgrp)+
	theme_bw(20)

knime-trans-snippet.R

library(ggplot2)
library(reshape2)
library(plyr)
library(dplyr)

knime.out<- knime.in %>% 
	select(Label,starts_with("CentroidDistance")) %>%
	ddply(.(Label),
	function(raw.df) 
		{
		in.df<-raw.df %>% melt(id.vars="Label")
		mutate(in.df,
			norm.value=value/sum(value),
			norm.vis=max(norm.value)-min(norm.value),
			std.vis=(max(value)-min(value))/mean(value)
		)[1,,drop=F]
		}) %>%
	subset(!is.na(norm.value)) %>%
	mutate(vis.class=cut_number(std.vis,4)) %>% 
	select(-variable,-value,-norm.value)