Tufte in R

Lukasz Piwek

• Updates
• Introduction
• Minimal line plot
• Range-frame (or quartile-frame) scatterplot
• Dot-dash (or rug) scatterplot
• Marginal histogram scatterplot
• Minimal boxplot
• Minimal barchart
• Slopegraph
• Sparklines
• Stem-and-leaf display
• Discussion

Updates

6th of July 2017: (1) New category - interactive plots made in Tufte-style with R - the first addition is basic line plot and basic barchart with the use of package highcharter; (2) Revised slopegraph in base graphics - Thomas Leeper has implemented his slopegraph functions into development version package on GitHub; (3) Revised sparklines in lattice get a gray bands (thanks to Bryan Urban for sharing his code on this); (4) Improved quality of all example figures; (5) Commented code and removed plot for range frame plot and dot-dash plot in base graphics produced with Steven Murdoch’s custom GitHub fancyaxis function - doesn’t work anymore; (6) Commented code and removed plot for slopegraph produced with James Keirstead’s GitHubs function - doesn’t work anymore.

30th of May 2016: (1) Added a new category - marginal histogram scatterplots for base graphics and ggplot2; (2) added new sparklines in base graphics with plotSparklineTable function from epanetReader package; (3) added new slopegraph in base graphic with bumpchart function from plotrix package (thanks to Jim Lemon forthis suggestion); (4) added range-frame scatterplot in ggplot2 with function qfplot by Mikhail Popov.

16th of February 2016: (1) Added stem-and-leaf displays; (2) added sparklines in ggplot2 (thanks to Wouter Van Der Bijl for addressing my StackOverflow question); (3) back to all-in-one-page format and revised sparklines to simplify; (4) range-frame and dot-dash plots split into separate sections for clarity; (5) overlong lines in ggplot2 version of dot-dash plots is back due to recent updates in ggplot2 that crash previous solution; (6) changed data set for slopegraph to match it between graphical systems.

17th of October 2015: Thank you for the kind words regarding this project! Major changes in this update: (1) added the first two methods to create Sparklines using Base Graphics and Lattice; (2) the document has become too large to keep on one-page. Its now split into two separate pages-chapters: Chapter 1: Line plot, Boxplot, Barchart and Slopegraph and Chapter 2: Sparklines; (3) corrected overlong lines produces in y-axis when making dot-dash plots with panel.rug() for Lattice (thanks to Josh O’Brien) and with geom_rug() for ggplot2 (thanks to BondedDust); (4) added this list of updates to keep a reasonable track of changes.

28th of July 2015: Tufte in R is live!

Introduction

Motivation

The idea behind Tufte in R is to use R - the most powerful open-source statistical programming language - to replicate excellent visualisation practices developed by Edward Tufte. It’s not a novel approach - there are plenty of excellent R functions and related packages wrote by people who have much more expertise in programming than myself. I simply collect those resources in one place in an accessible and replicable format, adding a few bits of my own coding discoveries.

Format

Each visualisation is provided in three graphical systems used in R: base graphics, lattice and ggplot2. As an example data I mainly use basic data sets easily accessible within R. Occasionally I use data from package psych developed by William Revelle, package MASS developed by Brian Ripley with collegues and various custom data I link via my Gist profile.

This page was produced in RMarkdown using Michael Sachs’s tuftehandout, but with a modified CSS inspired by Dave Liepmann’s Tufte CSS. Its best if you view this page on a desktop computer rather than mobile devices.

Requirements

You need the most recent version of R installed on you computer. You also need a basic understanding of R and there are some great online tutorials to get you started. I also recommend R Studio as an integrated development environment for R.

I use resources from a number of R packages. You can install all those packages at once via R console using the command below:

install.packages(c("CarletonStats", "devtools", "epanetReader", "fmsb", "ggplot2", "ggthemes", 
                   "latticeExtra", "MASS", "PerformanceAnalytics", "psych", 
                   "plyr", "prettyR", "plotrix", "proto", "RCurl", "reshape", "reshape2"))

Minimal line plot

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 65

We start by plotting the most basic graph from page 65 of The Visual Display of Quantitative Information - a minimal line plot. This one is important because it illustrates the most elemental principle - that of minimalism with reduced ‘data-ink’. As Tufte explains, the ‘data-ink’ (total ink used to print the graphic) ratio should equal to ‘1 - proportion of graphic that can be erased without loss of data-information’. The primary challenge is therefore to modify the default graphs produced with R so that we remove as much of ‘non-data ink’ as possible. As you will soon see, this is done by subtracting and deconstructing existing R graphs to get rid of as much ‘non-data ink’ as possible.

Minimal line plot in base graphics

Parameter axis = F prevents from drawing all axes elements so they can be easily refined with axis() function. I use minimal and maximal values from the data to draw text() - it usually requires a bit of tweaking to get it right. Font is changed to serif with family.

x <- 1967:1977
y <- c(0.5,1.8,4.6,5.3,5.3,5.7,5.4,5,5.5,6,5)
pdf(width=10, height=6)
plot(y ~ x, axes=F, xlab="", ylab="", pch=16, type="b")
axis(1, at=x, label=x, tick=F, family="serif")
axis(2, at=seq(1,6,1), label=sprintf("$%s", seq(300,400,20)), tick=F, las=2, family="serif")
abline(h=6,lty=2)
abline(h=5,lty=2)
text(max(x), min(y)*2.5,"Per capita\nbudget expanditures\nin constant dollars", adj=1, 
     family="serif")
text(max(x), max(y)/1.08, labels="5%", family="serif")
dev.off()

Minimal line plot in lattice

Arguments scales and par.settings have to be used heavily to customise scales and get rid of box. I used benbarnes axis hack from Stackoverflow to draw only axes ticks.

library(lattice)
x <- 1967:1977
y <- c(0.5,1.8,4.6,5.3,5.3,5.7,5.4,5,5.5,6,5)
xyplot(y~x, xlab="", ylab="", pch=16, col=1, border = "transparent", type="o",
       abline=list(h = c(max(y),max(y)-1), lty = 2),
       scales=list(x=list(at=x,labels=x, fontfamily="serif", cex=1),
                   y=list(at=seq(1,6,1), fontfamily="serif", cex=1,
                          label=sprintf("$%s",seq(300,400,20)))),
       par.settings = list(axis.line = list(col = "transparent"), dot.line=list(lwd=0)),
       axis = function(side, line.col = "black", ...) {
         if(side %in% c("left","bottom")) {axis.default(side = side, line.col = "black", ...)}})
ltext(current.panel.limits()$xlim[2]/1.1, adj=1, fontfamily="serif", 
      current.panel.limits()$ylim[1]/1.3, cex=1,
      "Per capita\nbudget expandures\nin constant dollars")
ltext(current.panel.limits()$xlim[2]/1.1, adj=1, fontfamily="serif", 
      current.panel.limits()$ylim[1]/5.5, cex=1, "5%")

Minimal line plot in ggplot2

I use excellent package ggthemes by Jeffrey B. Arnold which provides a lot of useful functions for Tufte-like plots - including a dedicated theme_tufte() function.

library(ggplot2)
library(ggthemes)
x <- 1967:1977
y <- c(0.5,1.8,4.6,5.3,5.3,5.7,5.4,5,5.5,6,5)
d <- data.frame(x, y)
ggplot(d, aes(x,y)) + geom_line() + geom_point(size=3) + theme_tufte(base_size = 15) +
  theme(axis.title=element_blank()) + geom_hline(yintercept = c(5,6), lty=2) + 
  scale_y_continuous(breaks=seq(1, 6, 1), label=sprintf("$%s",seq(300,400,20))) + 
  scale_x_continuous(breaks=x,label=x) +
  annotate("text", x = c(1977,1977.2), y = c(1.5,5.5), adj=1,  family="serif",
           label = c("Per capita\nbudget expandures\nin constant dollars", "5%"))

Minimal line plot - interactive plot with highcharter

A new approach to create a dynamic plots with package highcharter - seems its based on Java wrappers and includes dedicated hc_theme_tufte().

library(highcharter)
x <- 1967:1977
y <- c(290,318,372,385,385,372,386,380,390,400,380)
d <- data.frame(x, y)
highchart() %>%
  hc_chart(type = "scatter") %>% 
  hc_subtitle(text = "Per capita budget expanditures in constant dollars") %>%
  hc_yAxis(labels = list(format = "${value}")) %>%
  hc_add_series(data = d) %>% 
  hc_add_theme(hc_theme_tufte())

Range-frame (or quartile-frame) scatterplot

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 130-133. This doesn’t really replicate Tufte range frame because its a bit tricky to draw custom axis lines in basic graphics. As a rough starting point I use summary() to display values for minimum, maximum, median, mean and both quartiles on the axes.

Range frame plot in base graphics

x <- mtcars$wt
y <- mtcars$mpg
plot(x, y, main="", axes=FALSE, pch=16, cex=0.8, family="serif",
xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel")
axis(1,at=summary(x),labels=round(summary(x),1), tick=F, family="serif")
axis(2,at=summary(y),labels=round(summary(y),1), tick=F, las=2, family="serif")

Range frame plot in base graphics with fancyaxis

Steven Murdoch’s GitHub fancyaxis function - currently doesn’t work.

# library(devtools)
# source_url("https://raw.githubusercontent.com/sjmurdoch/fancyaxis/master/fancyaxis.R")
# x <- mtcars$wt
# y <- mtcars$mpg
# plot(x, y, main="", axes=FALSE, pch=16, cex=0.8,
#      xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel")
# fancyaxis(1, summary(x), digits=1)
# fancyaxis(2, summary(y), digits=1)

Range frame plot in lattice

Again, I used benbarnes axis hack from Stackoverflow to draw only axes ticks. Heavy use of par.settings to change the fontfamily to serif.

library(lattice)
x <- mtcars$wt
y <- mtcars$mpg
xyplot(y ~ x, mtcars, col=1, pch=16, fontfamily="serif",
xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel",
par.settings = list(axis.line = list(col="transparent"),
par.xlab.text=list(fontfamily="serif"),
par.ylab.text=list(fontfamily="serif")),
scales = list(x=list(at=summary(mtcars$wt),labels=round(summary(mtcars$wt),1),
fontfamily="serif"),
y=list(at=summary(mtcars$mpg),labels=round(summary(mtcars$mpg),1),
fontfamily="serif")),
axis = function(side, line.col = "black", ...) {
if(side %in% c("left","bottom")) {axis.default(side = side, line.col = "black", ...)}})

Range-frame plot in ggplot2

Another use of package ggthemes by Jeffrey B. Arnold - this time for geom_rangeframe().

library(ggplot2)
library(ggthemes)
ggplot(mtcars, aes(wt, mpg)) + geom_point() + geom_rangeframe() + theme_tufte() +
xlab("Car weight (lb/1000)") + ylab("Miles per gallon of fuel") + 
theme(axis.title.x = element_text(vjust=-0.5), axis.title.y = element_text(vjust=1.5))

Range-frame plot in ggplot2 with qfplot

Function qfplot by Mikhail Y. Popov - limited customisation, but simple to deploy.

library(devtools)
source_url('https://raw.githubusercontent.com/bearloga/Quartile-frame-Scatterplot/master/qfplot.R')
qfplot(x=mtcars$wt, y=mtcars$mpg, xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel")

Dot-dash (or rug) scatterplot

Dot-dash plot in base graphics with fancyaxis

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 130-133. Steven Murdoch’s GitHub minimalrug function (currently doesn’t work).

# library(devtools)
# source_url("https://raw.githubusercontent.com/sjmurdoch/fancyaxis/master/fancyaxis.R")
# x <- mtcars$wt
# y <- mtcars$mpg
# plot(x, y, main="", axes=FALSE, pch=16, cex=0.8,
# xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel", 
# xlim=c(min(x)-0.2, max(x)+0.2),
# ylim=c(min(y)-1.5, max(y)+1.5))
# axis(1, tick=F)
# axis(2, tick=F, las=2)
# minimalrug(x, side=1, line=-0.8)
# minimalrug(y, side=2, line=-0.8)

Dot-dash plot in lattice

A useful panel.rug() lattice function used to create a dot-dash axis with a neat solution from Josh O’Brien to control the length of dash margin lines.

library(lattice)
x <- mtcars$wt
y <- mtcars$mpg
xyplot(y ~ x, xlab="Car weight (lb/1000)", ylab="Miles per gallon of fuel",
par.settings = list(axis.line = list(col="transparent")),
panel = function(x, y,...) { 
panel.xyplot(x, y, col=1, pch=16)
panel.rug(x, y, col=1, x.units = rep("snpc", 2), y.units = rep("snpc", 2), ...)})

Dot-dash plot in ggplot2

Here I use a geom_rug() function from ggplot2.

library(ggplot2)
library(ggthemes)
ggplot(mtcars, aes(wt, mpg)) + geom_point() + geom_rug() + theme_tufte(ticks=F) + 
  xlab("Car weight (lb/1000)") + ylab("Miles per gallon of fuel") + 
  theme(axis.title.x = element_text(vjust=-0.5), axis.title.y = element_text(vjust=1))

Marginal histogram scatterplot

Marginal histogram scatterplot in base graphics with fancyaxis

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 133.
Steven Murdoch’s GitHub axisstripchart creates a marginal dot-histograms, although with limited customisation.

library(devtools)
source_url("https://raw.githubusercontent.com/sjmurdoch/fancyaxis/master/fancyaxis.R")
x <- faithful$waiting
y <- faithful$eruptions
plot(x, y, main="", axes=FALSE, pch=16, cex=0.8,
     xlab="Time till next eruption (min)", ylab="Duration (sec)", 
     xlim=c(min(x)/1.1, max(x)), ylim=c(min(y)/1.5, max(y)))
axis(1, tick=F)
axis(2, tick=F, las=2)
axisstripchart(faithful$waiting, 1)
axisstripchart(faithful$eruptions, 2)

Marginal histogram scatterplot in lattice (in preparation)

Marginal histogram scatterplot in ggplot2 with ggMarginal

Dean Attali’s ggExtra package includes function ggMarginal to create margin histograms. There is no option to create dot-histograms, just the standard bar-based histograms.

library(ggplot2)
library(ggExtra)
library(ggthemes)
p <- ggplot(faithful, aes(waiting, eruptions)) + geom_point() + theme_tufte(ticks=F)
ggMarginal(p, type = "histogram", fill="transparent")

However, ggMarginal can be also used to quickly create margin densityplots using the same function:

library(ggplot2)
library(ggExtra)
library(ggthemes)
p <- ggplot(faithful, aes(waiting, eruptions)) + geom_point() + theme_tufte(ticks=F) +
  theme(axis.title=element_blank(), axis.text=element_blank())
ggMarginal(p, type = "density")

Margin densityplot

…and it can also be used to create margin boxplots:

library(ggplot2)
library(ggExtra)
library(ggthemes)
p <- ggplot(faithful, aes(waiting, eruptions)) + geom_point() + theme_tufte(ticks=F) +
  theme(axis.title=element_blank(), axis.text=element_blank())
ggMarginal(p, type = "boxplot", size=10, fill="transparent")

Margin boxplot

Minimal boxplot

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 125 & 129. Argument pars is used to deconstruct the default base graphics boxplot.

Minimal boxplot in base graphics

x <- quakes$mag
y <- quakes$stations
boxplot(y ~ x, main = "", axes = FALSE, xlab=" ", ylab=" ",
pars = list(boxcol = "transparent", medlty = "blank", medpch=16, whisklty = c(1, 1),
medcex = 0.7,  outcex = 0, staplelty = "blank"))
axis(1, at=1:length(unique(x)), label=sort(unique(x)), tick=F, family="serif")
axis(2, las=2, tick=F, family="serif")
text(min(x)/3, max(y)/1.1, pos = 4, family="serif",
"Number of stations \nreporting Richter Magnitude\nof Fiji earthquakes (n=1000)")

Minimal boxplot in base graphics with chart.Boxplot

This uses chart.Boxplot function from PerformanceAnalytics package with dedicated as.Tufte=T argument. This solution requires data to be in wide table format and it has a limited customisation options.

library(PerformanceAnalytics)
library(psych)
d <- msq[,80:84]
chart.Boxplot(d, main = "", xlab="average personality rating (based on n=3896)", ylab="", 
element.color = "transparent", as.Tufte=TRUE)

Minimal boxplot in lattice

Argument par.settings is used to deconstruct the default lattice boxplot.

x <- quakes$mag
y <- quakes$stations
bwplot(y ~ x, horizontal=F, xlab="", ylab="", do.out = FALSE, box.ratio = 0,
scales=list(x=list(labels=sort(unique(x)), fontfamily="serif"),
y=list(fontfamily="serif")),
par.settings = list(axis.line = list(col = "transparent"), box.umbrella=list(lty=1, col= 1),
box.dot=list(col= 1), box.rectangle = list(col= c("transparent"))))
ltext(current.panel.limits()$xlim[1]+250, adj=1,
current.panel.limits()$ylim[2]+50, fontfamily="serif",
"Number of stations \nreporting Richter Magnitude\nof Fiji earthquakes (n=1000)")

Minimal boxplot in ggplot2

Function geom_tufteboxplot from package ggthemes is used to draw boxplot in ggplot2.

library(ggplot2)
library(ggthemes)
ggplot(quakes, aes(factor(mag),stations)) + theme_tufte() +
geom_tufteboxplot(outlier.colour="transparent") + theme(axis.title=element_blank()) +
annotate("text", x = 8, y = 120, adj=1,  family="serif",
label = c("Number of stations \nreporting Richter Magnitude\nof Fiji earthquakes (n=1000)"))

Minimal barchart

Minimal barchart in base graphics

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 125 & 129. Basic graphics has an awkward way to change the width of bars in barplot - especially when you want to draw axis names separately. It requires some tweaking of arguments width and space, as well as location in axis() function. I use ablines to draw Tufte-like grid lines.

library(psych)
d <- colMeans(msq[,c(2,7,34,36,42,43,46,55,68)], na.rm = T)*10
barplot(d, xaxt="n", yaxt="n", ylab="", border=F, width=c(.35), space=1.8)
axis(1, at=(1:length(d))-.26, labels=names(d), tick=F, family="serif")
axis(2, at=seq(1, 5, 1), las=2, tick=F, family="serif")
abline(h=seq(1, 5, 1), col="white", lwd=3)
abline(h=0, col="gray", lwd=2)
text(min(d)/2, max(d)/1.2, pos = 4, family="serif",
"Average scores\non negative emotion traits\nfrom 3896 participants\n(Watson et al., 1988)")

Minimal barchart in lattice

Lattice barchart draws bars horizontally by default and it gets messy if you change it to vertical bars. Function panel.abline is used to draw grid lines.

library(lattice)
library(psych)
d <- colMeans(msq[,c(2,7,34,36,42,43,46,55,68)],na.rm = T)*10
barchart(sort(d), xlab="", ylab="", col = "grey", origin=1,  
border = "transparent", box.ratio=0.5, 
panel = function(x,y,...) {
panel.barchart(x,y,...)
panel.abline(v=seq(1,6,1), col="white", lwd=3)},
par.settings = list(axis.line = list(col = "transparent")))
ltext(current.panel.limits()$xlim[2]-50, adj=1,  
current.panel.limits()$ylim[1]-100,
"Average scores\non negative emotion traits\nfrom 3896 participants\n(Watson et al., 1988)")

Minimal barchart in ggplot2

library(ggplot2)
library(ggthemes)
library(psych)
library(reshape2)
d <- melt(colMeans(msq[,c(2,7,34,36,42,43,46,55,68)],na.rm = T)*10)
d$trait <- rownames(d)
ggplot(d, aes(x=trait, y=value)) + theme_tufte(base_size=14, ticks=F) +
  geom_bar(width=0.25, fill="gray", stat = "identity") +  theme(axis.title=element_blank()) +
  scale_y_continuous(breaks=seq(1, 5, 1)) + 
  geom_hline(yintercept=seq(1, 5, 1), col="white", lwd=1) +
  annotate("text", x = 3.5, y = 5, adj=1,  family="serif",
label = c("Average scores\non negative emotion traits
          from 3896 participants\n(Watson et al., 1988)"))

Minimal barchart - interactive with highcharter

library(psych)
library(reshape)
library(highcharter)
values <- 1 + abs(rnorm(12))
d <- melt(colMeans(msq[,c(2,7,34,36,42,43,46,55,68)], na.rm = T)*10)
trait <- row.names(d) 
value <- as.vector(d[,1])
highchart() %>%
  hc_chart(type = "column") %>%
  hc_add_series(data = value) %>%
  hc_xAxis(categories = row.names(d)) %>%
  hc_add_theme(hc_theme_tufte2())

Slopegraph

Slopegraph in base graphics

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 158. Using Thomas Leeper’s slopegraph package.

The most promising slopegraph functions for base graphics and ggplot2 comes from Thomas Leeper slopegraph package. Thomas’s solutions have evolved gradually and it’s now the most efficient method to create slopegraphs in R. However, a major limitation is inability to efficently offset left and right side labels to avoid don’t overlap (as seen below).

library(devtools)
#install_github("leeper/slopegraph")#install Leeper's package from Github
library(slopegraph)
data(cancer)
slopegraph(cancer, col.lines = 'gray', col.lab = 1, col.num = 1,
           xlim = c(-.2,5),
           main = "Estimate of % survival rates",
           xlabels = c('5 Year','10 Year','15 Year','20 Year'))

Slopegraph in lattice (might not happen)

Issue with ggslopegraph has been logged here on Github

Slopegraph in ggplot2 with ggslopegraph (with bugs, in preparation)

Slopegraph in ggplot2 with plot_slopegraph

James Keirstead’s GitHubs function (currently doesn’t work).

# library(ggplot2)
# library(ggthemes)
# library(devtools)
# library(RCurl)
# library(plyr)
# source_url("https://raw.githubusercontent.com/jkeirstead/r-slopegraph/master/slopegraph.r")
# d <- read.csv(text = getURL("https://raw.githubusercontent.com/jkeirstead/r-slopegraph/master/cancer_survival_rates.csv"))
# df <- build_slopegraph(d, x="year", y="value", group="group", method="tufte", min.space=0.04)
# df <- transform(df, x=factor(x, levels=c(5,10,15,20),
#                              labels=c("5 years","10 years","15 years","20 years")), y=round(y))
# plot_slopegraph(df) + labs(title="Estimates of % survival rates") +
#   theme_tufte(base_size=16, ticks=F) + theme(axis.title=element_blank())

Sparklines

There is no ‘out-of-box’ solution in the existing packages that truly replicate Tufte-style sparklines. Main issues are scaling the size of the plot and labeling of the points - those factors are likely to change depending on the data set you’re plotting, so you will have to adjust specific parameters (which I highlight for every graphical system). To make the output more consistent, every sparkline plot will be automatically saved in the working directory in a vector format as a PDF (using pdf() and dev.off() functions).

A word of warning - in its current format, making sparklines requires a bit more advanced knowledge of R. Its far from perfect - proceed with caution.

Sparklines in base graphics

Sparklines in base graphics use some elements of functions from YaleToolkit developed by John Emerson and Walton Green. In particular, it’s a result of mine and Ben’s hacking of YaleToolkit functions on Stackoverflow. I’ve use a simple loop that takes a number of columns in a data set and creates as much sparklines as there are columns. In the same manner I use mfrow parameter in par() function to set the number of rows to a number of columns in data frame.

Data on US Rate of Crime (per 100,000 people) from 1960 to 2014 comes from FBI UCS Annual Crime Reports compiled by disastercenter.com.

library(RCurl)
dd <- read.csv(text = getURL("https://gist.githubusercontent.com/GeekOnAcid/da022affd36310c96cd4/raw/9c2ac2b033979fcf14a8d9b2e3e390a4bcc6f0e3/us_nr_of_crimes_1960_2014.csv"))
d <- dd[,c(2:11)]
pdf("sparklines_base.pdf", height=10, width=6)
par(mfrow=c(ncol(d),1), mar=c(1,0,0,8), oma=c(4,1,4,4))
for (i in 1:ncol(d)){
  plot(d[,i], lwd=0.5, axes=F, ylab="", xlab="", main="", type="l", new=F)
  axis(4, at=d[nrow(d),i], labels=round(d[nrow(d),i]), tick=F, las=1, line=-1.5, 
       family="serif", cex.axis=1.2)
  axis(4, at=d[nrow(d),i], labels=names(d[i]), tick=F, line=1.5, 
       family="serif", cex.axis=1.4, las=1)
  text(which.max(d[,i]), max(d[,i]), labels=round(max(d[,i]),0), 
       family="serif", cex=1.2, adj=c(0.5,3))
  text(which.min(d[,i]), min(d[,i]), labels=round(min(d[,i]),0), 
       family="serif", cex=1.2, adj=c(0.5,-2.5))
  ymin <- min(d[,i]); tmin <- which.min(d[,i]); ymax<-max(d[,i]); tmax<-which.max(d[,i]);
  points(x=c(tmin,tmax), y=c(ymin,ymax), pch=19, col=c("red","blue"), cex=1)
  rect(0, summary(d[,i])[2], nrow(d), summary(d[,i])[4], border=0, 
       col = rgb(190, 190, 190, alpha=90, maxColorValue=255))}
axis(1, at=1:nrow(dd), labels=dd$Year, pos=c(-5), tick=F, family="serif", cex.axis=1.4)
dev.off()

Sparklines in base graphics with plotSparklineTable

This uses plotSparklineTable function from epanetReader package by Bradley Eck. This compact solution requires data to be in long table format and it has a limited customisation options. Great for making a rapid summaries with sparklines.

library(epanetReader)
library(reshape)
library(RCurl)
dd <- read.csv(text = getURL("https://gist.githubusercontent.com/GeekOnAcid/da022affd36310c96cd4/raw/9c2ac2b033979fcf14a8d9b2e3e390a4bcc6f0e3/us_nr_of_crimes_1960_2014.csv"))
d <- melt(dd[,c(2:11)])
pdf("sparklines_base_epanetReader.pdf", height=6, width=10)
plotSparklineTable(d, row.var = 'variable', col.vars = 'value')
dev.off()

Sparklines in lattice

You have much better control over the location and size of sparklines when you use lattice. The only problem are right-side labels for which you have to use grid library in order to ‘hack’ the view parameters with functions pushViewport() and popViewport(). You can learn more about this in an extensive collection of grid vignettes.

library(lattice)
library(latticeExtra)
library(grid)
library(reshape)
library(RCurl)
dd <- read.csv(text = getURL("https://gist.githubusercontent.com/GeekOnAcid/da022affd36310c96cd4/raw/9c2ac2b033979fcf14a8d9b2e3e390a4bcc6f0e3/us_nr_of_crimes_1960_2014.csv"))
d <- melt(dd, id="Year")
names(d)[1] <- "time"
pdf("sparklines_lattice.pdf", height=10, width=8)
xyplot(value~time | variable, d, xlab="", ylab="", strip=F, lwd=0.7, col=1, type="l",
       layout=c(1,length(unique(d$variable))), between = list(y = 1),
       scales=list(y=list(at=NULL, relation="free"), x=list(fontfamily="serif")),
       par.settings = list(axis.line = list(col = "transparent"),
                           layout.widths=list(right.padding=20, left.padding=-5)),
       panel = function(x, y, ...) {
         panel.xyplot(x, y, ...)
         pushViewport(viewport(xscale=current.viewport()$xscale-5,
                               yscale=current.viewport()$yscale, clip="off"))
         panel.text(x=tail(x,n=1), y=tail(y,n=1), labels=levels(d$variable)[panel.number()],
                    fontfamily="serif", pos=4)
         popViewport()
         panel.text(x=x[which.max(y)], y=max(y), labels=round(max(y),0), cex=0.8,
                    fontfamily="serif",adj=c(0.5,2.5))
         panel.text(x=x[which.min(y)], y=min(y), labels=round(min(y),0), cex=0.8,
                    fontfamily="serif",adj=c(0.5,-1.5))
         panel.text(x=tail(x,n=1), y=tail(y,n=1), labels=round(tail(y,n=1),0), cex=0.8,
                    fontfamily="serif", pos=4)
         panel.points(x[which.max(y)], max(y),  pch=16, cex=1)
         panel.points(x[which.min(y)], min(y),  pch=16, cex=1, col="red")
         panel.rect(min(x), quantile(y, 0.25), max(x), quantile(y, 0.75),
                    col = "grey", border = "transparent", alpha = 0.4)
       })
dev.off()

Sparklines in ggplot2

Created with the help of Wouter Van Der Bijl as part of this StackOverflow question.

library(ggplot2)
library(ggthemes)
library(dplyr)
library(reshape)
library(RCurl)
dd <- read.csv(text = getURL("https://gist.githubusercontent.com/GeekOnAcid/da022affd36310c96cd4/raw/9c2ac2b033979fcf14a8d9b2e3e390a4bcc6f0e3/us_nr_of_crimes_1960_2014.csv"))
d <- melt(dd, id="Year")
names(d) <- c("Year","Crime.Type","Crime.Rate")
d$Crime.Rate <- round(d$Crime.Rate,0)
mins <- group_by(d, Crime.Type) %>% slice(which.min(Crime.Rate))
maxs <- group_by(d, Crime.Type) %>% slice(which.max(Crime.Rate))
ends <- group_by(d, Crime.Type) %>% filter(Year == max(Year))
quarts <- d %>% group_by(Crime.Type) %>%
  summarize(quart1 = quantile(Crime.Rate, 0.25),
            quart2 = quantile(Crime.Rate, 0.75)) %>%
  right_join(d)
pdf("sparklines_ggplot.pdf", height=10, width=8)
ggplot(d, aes(x=Year, y=Crime.Rate)) + 
  facet_grid(Crime.Type ~ ., scales = "free_y") + 
  geom_ribbon(data = quarts, aes(ymin = quart1, max = quart2), fill = 'grey90') +
  geom_line(size=0.3) +
  geom_point(data = mins, col = 'red') +
  geom_point(data = maxs, col = 'blue') +
  geom_text(data = mins, aes(label = Crime.Rate), vjust = -1) +
  geom_text(data = maxs, aes(label = Crime.Rate), vjust = 2.5) +
  geom_text(data = ends, aes(label = Crime.Rate), hjust = 0, nudge_x = 1) +
  geom_text(data = ends, aes(label = Crime.Type), hjust = 0, nudge_x = 5) +
  expand_limits(x = max(d$Year) + (0.25 * (max(d$Year) - min(d$Year)))) +
  scale_x_continuous(breaks = seq(1960, 2010, 10)) +
  scale_y_continuous(expand = c(0.1, 0)) +
  theme_tufte(base_size = 15, base_family = "Helvetica") +
  theme(axis.title=element_blank(), axis.text.y = element_blank(), 
        axis.ticks = element_blank(), strip.text = element_blank())
dev.off()

Stem-and-leaf display

Edward Tufte, The Visual Display of Quantitative Information (Cheshire, 1983), p. 140 Based on John Tukey, Exploratory Data Analysis (Addison-Wesley, 1970).

Stem-and-leaf display is not exactly a ‘Tuftesque’ solution as it invented in the beginning of 20 century but was only popularised in 1980s by John Tukey. A stem-and-leaf display is a display for presenting quantitative data in a graphical format, similar to a histogram, to assist in visualizing the shape of a distribution. Stem-and-leaf plot is the only visualisation in this collection thats printed in the console in R rather than being processed with any graphical system.

Stem-and-leaf display in console with base graphics

Data shows the duration of the eruption for the Old Faithful geyser in Yellowstone National Park, Wyoming, USA.

stem(faithful$eruptions)

## 
##   The decimal point is 1 digit(s) to the left of the |
## 
##   16 | 070355555588
##   18 | 000022233333335577777777888822335777888
##   20 | 00002223378800035778
##   22 | 0002335578023578
##   24 | 00228
##   26 | 23
##   28 | 080
##   30 | 7
##   32 | 2337
##   34 | 250077
##   36 | 0000823577
##   38 | 2333335582225577
##   40 | 0000003357788888002233555577778
##   42 | 03335555778800233333555577778
##   44 | 02222335557780000000023333357778888
##   46 | 0000233357700000023578
##   48 | 00000022335800333
##   50 | 0370

Stem-and-leaf display in console with CarletonStats

Data set taken from package MASS. It shows risk factors associated with low infant birth weight. The data were collected at Baystate Medical Center, Springfield, Mass during 1986.

The stemPlot function expands the basic stem plot by accepting a factor variable as a second argument to create stem plots for each of the levels.

library(CarletonStats)
library(MASS)
stemPlot(birthwt$bwt, birthwt$smoke, varname="infant birth weight (in grams)",
         grpvarname="whether mother smoked during pregnancy (1) or not (0)")

## 
## ***Stem and Leaf plot for  infant birth weight (in grams) ***
##    Grouped by levels of  whether mother smoked during pregnancy (1) or not (0) 
## 
##     0 
##  :
##   The decimal point is 2 digit(s) to the right of the |
## 
##   10 | 2
##   12 | 3
##   14 | 799
##   16 | 03
##   18 | 9037
##   20 | 66809
##   22 | 4480358
##   24 | 14450025
##   26 | 24423558
##   28 | 144468822288
##   30 | 6668990088
##   32 | 003333377227
##   34 | 0266794479
##   36 | 011355037779
##   38 | 0366814478
##   40 | 00551577
##   42 | 
##   44 | 9
##   46 | 
##   48 | 9
## 
## 
##     1 
##  :
##   The decimal point is 3 digit(s) to the right of the |
## 
##   0 | 7
##   1 | 1
##   1 | 889999
##   2 | 11112223344444444
##   2 | 5555566677888899999
##   3 | 0000011111233333444
##   3 | 6666778999
##   4 | 2

Stem-and-leaf display in base graphics with fmsb

A nice function gstem from package fmsb by Minato Nakazawa allows to wrap the console output of stem-and-leaf display in a base graphics wrapper.

library(fmsb)
gstem(faithful$eruptions)

Discussion