The animation pushes the user to use his/her memory in order to compare the data for each company. On the other hand, on the stacked barcharts we have lower cognitive load and the comparison can be done directly. Moreover the use of the length as a channel (not area, since width is constant for all of bars) helps the user to perceive the actual data more accuratelly than the angle channel in the piecharts.
idea : http://www.storytellingwithdata.com/blog/2014/08/design-with-audience-in-mind
original data : http://www.usatoday.com/story/tech/2014/08/12/apple-diversity-white-asian-male/13951329/
dat <- read.table(text = " Apple Yahoo Google Facebook LinkedIn Twitter Ebay
1 70 62 70 69 61 70 58
2 30 37 30 31 39 30 42
3 0 1 0 0 0 0 0",header = T)
#make room for the legend
par(xpd=T, mar=par()$mar+c(0,0,0,8))
#make barchart with no y axis
barplot(as.matrix(dat),col=rainbow(3),main="Workforce by Gender",yaxt="n")
#add a y-axis with percentages
axis(2, at=pretty(a), lab=paste0(pretty(a),"%"), las=TRUE)
#legend
labels <- c("Male","Female","Not Disclosed")
legend(8.5, 80,labels,fill=rainbow(3),cex=0.8);
Live at /christost/a679815370ad5e8ce609
forked from ChristosT's block: CS 725 Information Visualization - VI4
https://cdnjs.cloudflare.com/ajax/libs/d3/3.5.5/d3.min.js