2

I want to use the nPlot() function in RCharts to plot percentages of people falling into discrete groups, rather than frequencies.

For example: Using the HairEyeColor data/code below I am able to consider the percentage of people with different hair colors (my grouping variable), as a function of their eye color. 

## server.r
library(rCharts)
library(shiny)
library(reshape)

HairEyeColor1 <- melt(round(prop.table(HairEyeColor[,,1],2)*100,2))
names(HairEyeColor1) <- c("Hair", "Eye", "Percent")

shinyServer(function(input, output) {
output$myChart <- renderChart({
p1 <- nPlot(Percent ~ Eye, group = "Hair", data = HairEyeColor1, type = multiBarChart")
p1$addParams(dom = 'myChart')
return(p1)
})
})

## ui.R
library(rCharts)
library(shiny)

shinyUI(pageWithSidebar(
headerPanel("rCharts: Interactive Charts from R using NVD3.js"),
sidebarPanel(
wellPanel(
    helpText(   "Look at the pretty graph"
    )
    ),
wellPanel(
    helpText(   "Look at the pretty graph"
    )
    ),
wellPanel(
    helpText(   "Look at the pretty graph"
    )
    )
),
mainPanel(
div(class='wrapper',
tags$style(".Nvd3{ height: 400px;}"),
showOutput("myChart","Nvd3")
)
)
))

Say I just wanted to look at a single factor, like Hair. Is there a way to plot their percentages using nPlot()?

prop.table(rowSums(HairEyeColor[,,1]))
    Black     Brown       Red       Blond 
    0.2007168 0.5125448   0.1218638 0.1648746

With type=multiBarChart I have tried:

p1 <- nPlot(Percent ~ , group = "Hair", data = HairEyeColor1, type = multiBarChart")

This fails entirely. I have also tried: 

p1 <- nPlot(Percent ~ 1, group = "Hair", data = HairEyeColor1, type = multiBarChart")

This at least passes some graph to Shiny UI. But this looks hideous and functionality (appearance of X/Y-axis, clicking on graph) is all lost.

I thought this would be appropriate for nPlot(type=discreteBarChart) but the data layout for this seems to want a dataframe with a single factor variable. So I can't quite see how to trick nPlot(type=discreteBarChart) into taking a vector of proportions/percentages.

Any suggestions appreciated.

Chris
  • 3,401
  • 5
  • 33
  • 42
  • 2
    Can you show the plot you are looking to make, made using base R, lattice or ggplot2? That will help me tell you how to make it with nPlot. – Ramnath Oct 31 '13 at 22:05
  • Was hoping for something like: `barplot(prop.table(rowSums(HairEyeColor[,,1])), ylim=c(0,1))` – Chris Nov 01 '13 at 00:13
  • I guess I wanted it to stack (into a single column) and have the height go to 100% as well. And have the option to remove certain levels by clicking factor level top right. But as I look into `discreteBarChart` more I realize this is not an option. And only available in `multiBarChart`. I guess like this too (with legend added obviously): `barplot(as.matrix(prop.table(rowSums(HairEyeColor[,,1]))), ylim=c(0,1),beside=FALSE)` – Chris Nov 01 '13 at 00:37

1 Answers1

2

Here is how to do it with nPlot. You can see the resulting plot here. If you want to stack the bars by default, add the line n1$chart(stacked = TRUE) before you print the chart.

# prepare data
require(plyr)
dat = as.data.frame(HairEyeColor)
dat = ddply(dat, .(Hair), summarize, Freq = sum(Freq), Group = "A")

# draw chart
require(rCharts)
n1 <- nPlot(Freq ~ Group, data = dat, group = 'Hair', type = 'multiBarChart')
n1
Ramnath
  • 54,439
  • 16
  • 125
  • 152
  • I tried `p1$print('inline', include_assets = F, cdn = FALSE)`, `p1$show('inline', include_assets = F, cdn = FALSE)`, `p1`, `p1$show()`, `p1$print()` but no one is workable to me. http://rpubs.com/englianhu/Milestone-Report similar with : 1) http://stackoverflow.com/questions/29029718/display-two-rcharts-nvd3-figures-next-to-each-other-in-rmarkdown 2) http://stackoverflow.com/questions/30212788/are-rcharts-and-dt-compatible-in-rmarkdown – Rγσ ξηg Lιαη Ημ 雷欧 Apr 25 '16 at 21:39