ggplot.Rnw

% Introduction to ggplot2
% Author: Karthik Ram, karthik.ram@gmail.com
% Licence, CC-BY
\documentclass{beamer}
\usepackage{listings}
\usepackage{inconsolata}
\setbeamertemplate{frametitle}[default][center]
\usepackage{url}
\setcounter{secnumdepth}{-1}
\usetheme{Amsterdam}
% --------------------------------------------------------------
% Setting up some knitr options
<<setup, include=FALSE>>=
opts_chunk$set(cache=TRUE, message=FALSE)
# smaller font size for chunks
opts_chunk$set(size = 'footnotesize')
options(width = 60)
@

% --------------------------------------------------------------
\begin{document}
\title{Data Visualization with R \& ggplot2}
\author{Karthik Ram}
\maketitle

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Download this PDF}
\begingroup
    \fontsize{12pt}{12pt}\selectfont
\href{http://github.com/karthikram/ggplot-lecture}{github.com/karthikram/ggplot-lecture}\\
\href{https://speakerdeck.com/karthik/}{https://speakerdeck.com/karthik/}
\endgroup
\includegraphics[scale=.31]{images/git_repo.png}
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Some housekeeping}
Install some packages (make sure you also have recent copies of reshape2 and plyr)
<<installation, tidy=FALSE, echo=TRUE, eval=FALSE>>=
install.packages("ggplot2", dependencies = TRUE)
@
\end{frame}



% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Base graphics}
\begin{itemize}
\item Ugly, laborious, and verbose\\
\item There are better ways to describe statistical visualizations.\\
\end{itemize}
\end{frame}

\begin{frame}[fragile]
\frametitle{Why \texttt{ggplot2}?}
\begin{itemize}
\item Follows a grammar, just like any language.
\item It defines basic components that make up a sentence. In this case, the grammar defines components in a plot.
\item Grammar of graphics originally coined by Lee Wilkinson
\end{itemize}
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Why \texttt{ggplot2}?}
\begin{itemize}
\item  Supports a continuum of expertise.
\item Get started right away but with practice you can effortless build complex, publication quality figures.
\end{itemize}
\end{frame}

% --------------------------------------------------------------
\section*{Basics}
\frame{\sectionpage}


\begin{frame}[fragile]
\frametitle{Some terminology}
\begin{itemize}
\item \textbf{ggplot} - The main function where you specify the dataset and variables to plot\\
\item \textbf{geoms} - geometric objects
    \begin{itemize}
    \item geom\_point(), geom\_bar(), geom\_density(), geom\_line(), geom\_area()
    \end{itemize}
\item \textbf{aes} -  aesthetics
        \begin{itemize}
    \item shape, transparency (alpha), color, fill, linetype.
    \end{itemize}
\item \textbf{scales}  Define how your data will be plotted
        \begin{itemize}
    \item \emph{continuous}, \emph{discrete}, \emph{log}
    \end{itemize}
\end{itemize}
\end{frame}


% --------------------------------------------------------------
\section*{Assembling your first ggplot}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{The iris dataset}
<<some_data, tidy=FALSE, echo=TRUE>>=
head(iris)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Let's try an example}
<<first_plot , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(data = iris, aes(x = Sepal.Length, y = Sepal.Width)) +
geom_point()
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Basic structure}
<<first_plotb , eval=FALSE, tidy=FALSE>>=
ggplot(data = iris, aes(x = Sepal.Length, y = Sepal.Width))
 + geom_point()
myplot <- ggplot(data = iris, aes(x = Sepal.Length, y = Sepal.Width))
myplot + geom_point()
 @
\begin{itemize}
\item Specify the data and variables inside the \texttt{ggplot} function.
\item Anything else that goes in here becomes a global setting.
\item Then add layers of geometric objects, statistical models, and panels.
\end{itemize}
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Quick note}
\begin{itemize}
\item Never use \texttt{qplot} - short for quick plot.
\item You'll end up unlearning and relearning a good bit.
\end{itemize}

\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Increase the size of points}
<<first_plot_size , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(data = iris, aes(x = Sepal.Length, y = Sepal.Width)) +
geom_point(size = 3)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Add some color}
<<first_plot_color , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(size = 3)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Differentiate points by shape}
<<first_plot_shape , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(aes(shape = Species), size = 3)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Exercise 1}
<<subset_data, tidy=FALSE, eval=FALSE>>=
# Make a small sample of the diamonds dataset
d2 <- diamonds[sample(1:dim(diamonds)[1], 1000), ]
@
Then generate this plot below.

<<ex1, echo = FALSE, out.width='.75\\linewidth', fig.height=3>>=
d2 <- diamonds[sample(1:dim(diamonds)[1], 1000), ]
ggplot(d2, aes(carat, price, color = color)) + geom_point() + theme_gray()
@
\end{frame}
% --------------------------------------------------------------
\section*{Box plots}
\frame{\sectionpage}

\begin{frame}[fragile]
See \texttt{?geom\_boxplot} for list of options
<<boxplots1 , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
library(MASS)
ggplot(birthwt, aes(factor(race), bwt)) + geom_boxplot()
@
\end{frame}


% --------------------------------------------------------------
\section*{Histograms}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
See \texttt{?geom\_histogram} for list of options
<<histogr , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
h <- ggplot(faithful, aes(x = waiting))
h + geom_histogram(binwidth = 30, colour = "black")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
<<histogra , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
h <- ggplot(faithful, aes(x = waiting))
h + geom_histogram(binwidth = 8, fill = "steelblue",
colour = "black")
@
\end{frame}

% --------------------------------------------------------------
\section*{Line plots}
\frame{\sectionpage}

\begin{frame}[fragile]
<<line_setup, echo=FALSE>>=
setwd('~/Github/ggplot2-lecture/')
@
% climate <- read.csv(text = RCurl::getURL('https://raw.github.com/karthikram/ggplot-lecture/master/climate.csv'))
<<linea , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
climate <- read.csv("climate.csv", header = T)
ggplot(climate, aes(Year, Anomaly10y)) +
geom_line()
@
\begin{flushright}
\begingroup
    \fontsize{6pt}{12pt}\selectfont
    \begin{verbatim}
        climate <- read.csv(text =
        RCurl::getURL('https://raw.github.com/karthikram/ggplot-lecture/master/climate.csv'))
    \end{verbatim}
\endgroup
\end{flushright}
\end{frame}

\begin{frame}[fragile]
We can also plot confidence regions
<<lineb , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(climate, aes(Year, Anomaly10y)) +
geom_ribbon(aes(ymin = Anomaly10y - Unc10y,
ymax = Anomaly10y + Unc10y),
fill = "blue", alpha = .1) +
geom_line(color = "steelblue")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Exercise 2}
\begin{itemize}
\item Modify the previous plot and change it such that there are three lines instead of one with a confidence band.
<<ex2, echo = FALSE, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold'>>=
cplot <- ggplot(climate, aes(Year, Anomaly10y))
cplot <- cplot + geom_line(size = 0.7, color = "black")
cplot <- cplot + geom_line(aes(Year, Anomaly10y + Unc10y), linetype = "dashed", size = 0.7, color = "red")
cplot <- cplot + geom_line(aes(Year, Anomaly10y - Unc10y), linetype = "dashed", size = 0.7, color = "red")
cplot + theme_gray()
@

\end{itemize}
\end{frame}


% --------------------------------------------------------------
\section*{Bar plots}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
<<barone , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(iris, aes(Species, Sepal.Length)) +
geom_bar(stat = "identity")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
<<bartwo , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
df  <- melt(iris, id.vars = "Species")
ggplot(df, aes(Species, value, fill = variable)) +
geom_bar(stat = "identity")
@
\end{frame}



% --------------------------------------------------------------
\section*{plyr and reshape are key for using \texttt{R}}
\frame{\sectionpage}

\begin{frame}[fragile]
\frametitle{plyr and reshape}
These two packages are the swiss army knives of R.
\begin{itemize}
\item \texttt{plyr}
    \begin{enumerate}
    \item ddply
    \item llply
    \item join
    \end{enumerate}
\item \texttt{reshape}.
    \begin{enumerate}
    \item melt
    \item dcast
    \item acast
    \end{enumerate}
\end{itemize}
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
<<bartwodata , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
iris[1:2, ]
df  <- melt(iris, id.vars = "Species")
df[1:2, ]
@
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
<<barthree , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(df, aes(Species, value, fill = variable)) +
geom_bar(stat = "identity", position = "dodge")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Exercise 3}
Using the d2 dataset you created earlier, generate this plot below. Take a quick look at the data first to see if it needs to be binned.
<<ex3, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold', echo = FALSE>>=
ggplot(d2, aes(clarity, fill = cut)) +
geom_bar(position = "dodge",stat = "bin") + theme_gray()
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Exercise 4}
\begin{itemize}
\item Using the climate dataset, create a new variable called sign. Make it logical (true/false) based on the sign of Anomaly10y.
\item Plot a bar plot and use \texttt{sign} variable as the fill.\\
<<ex4, echo = FALSE, warning = FALSE, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold'>>=
clim <- read.csv('climate.csv', header = TRUE)
clim$sign <- ifelse(clim$Anomaly10y<0, FALSE, TRUE)
# or as simple as
# clim$sign <- clim$Anomaly10y < 0
ggplot(clim, aes(Year, Anomaly10y)) + geom_bar(stat = "identity", aes(fill = sign)) + theme_gray()
@

\end{itemize}
\end{frame}


% --------------------------------------------------------------
\section*{Density Plots}
\frame{\sectionpage}

\begin{frame}[fragile]
\frametitle{Density plots}
<<densityone , eval=TRUE, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(faithful, aes(waiting)) + geom_density()
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Density plots}
<<densityonefove , eval=TRUE, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(faithful, aes(waiting)) +
geom_density(fill = "blue", alpha = 0.1)
@
\end{frame}



% --------------------------------------------------------------
\begin{frame}[fragile]
<<densitytwo , , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
ggplot(faithful, aes(waiting)) +
geom_line(stat = "density")
@
\end{frame}


% --------------------------------------------------------------
\section*{Mapping Variables to colors}
\frame{\sectionpage}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Colors}
<<color_list2, eval = FALSE, fig.width=4, fig.height=6, out.width='.75\\linewidth'>>=
# Map all points to one color
aes(color = "black")
# Or map the points to a variavble
aes(color = variable)
# Then add a scale for the colors.
# Below we manually define colors but there are other ways (see next slide)
scale_fill_manual(values = c("color1", "color2"))
@
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{The RColorBrewer package}
<<color_list, eval = FALSE, fig.width=4, fig.height=5, out.width='.75\\linewidth'>>=
library(RColorBrewer)
display.brewer.all()
@
\includegraphics[scale=0.25]{images/color_palette.png}
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Using a color brewer palette}
<<barcolors, fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
df  <- melt(iris, id.vars = "Species")
ggplot(df, aes(Species, value, fill = variable)) +
geom_bar(stat = "identity", position = "dodge") +
scale_fill_brewer(palette = "Set1")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Manual color scale}
<<facetgridcolors, eval = TRUE, fig.width=4, fig.height=3, out.width='.75\\linewidth', tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point() +
facet_grid(Species ~ .) +
scale_color_manual(values = c("red", "green", "blue"))
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Refer to a color chart for beautiful visualizations}
\url{http://tools.medialab.sciences-po.fr/iwanthue/}
\includegraphics[scale=0.25]{images/color_schemes.png}
\end{frame}

% -----------------------------------

\section*{Faceting}
\frame{\sectionpage}

\begin{frame}[fragile]
\frametitle{Faceting along columns}
<<facetgrid1, eval = TRUE, fig.width=4, fig.height=3, out.width='.75\\linewidth', tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point() +
facet_grid(Species ~ .)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{and along rows}
<<facet_grid2, eval = TRUE, fig.width=4, fig.height=3, out.width='.75\\linewidth', tidy= FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point() +
facet_grid(. ~ Species)
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{or just wrap your panels}
<<facet_wrap, eval = TRUE, fig.width=4, fig.height=3, out.width='.75\\linewidth', tidy = FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point() +
facet_wrap( ~ Species)
@
\end{frame}

% --------------------------------------------------------------
\section*{Adding smoothers}
\frame{\sectionpage}


% --------------------------------------------------------------
\begin{frame}[fragile]
<<adding_stats,, eval = TRUE, fig.width=5, fig.height=2, out.width='.75\\linewidth', tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(aes(shape = Species), size = 3) +
geom_smooth(method = "lm")
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
<<adding_stats2,, eval = TRUE, tidy=FALSE, fig.width=7, fig.height=3, out.width='.75\\linewidth'>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(aes(shape = Species), size = 3) +
geom_smooth(method = "lm") +
facet_grid(. ~ Species)
@
\end{frame}

% --------------------------------------------------------------
\section*{Themes}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Adding themes}
Themes are a great way to define custom plots.
<<theme_list, eval = FALSE, fig.width=4, fig.height=6, out.width='.75\\linewidth'>>=
+ theme()
# see ?theme() for more options
@

\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{A themed plot}
<<facet_wrap_theme, eval = FALSE, fig.width=4, fig.height=6, out.width='.75\\linewidth', tidy=FALSE>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(size = 1.2, shape = 16) +
facet_wrap( ~ Species) +
theme(legend.key = element_rect(fill = NA),
legend.position = "bottom",
strip.background = element_rect(fill = NA),
axis.title.y = element_text(angle = 0))
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Adding themes}
<<facet_wrap_theme_execc, eval = TRUE, echo = FALSE, fig.width=4, fig.height=3, out.width='.75\\linewidth'>>=
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(size = 1.2, shape = 16) +
facet_wrap( ~ Species) +
theme(legend.key = element_rect(fill = NA),
legend.position = "bottom",
strip.background = element_rect(fill = NA),
axis.title.y = element_text(angle = 0))
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{ggthemes library}
<<facet_wrap_theme_exec, eval = FALSE, echo = TRUE, fig.width=4, fig.height=3, out.width='.75\\linewidth'>>=
install.packages('ggthemes')
library(ggthemes)
# Then add one of these themes to your plot
 + theme_stata()
 + theme_excel()
 + theme_wsj()
 + theme_solarized()
@
\end{frame}


% --------------------------------------------------------------
\section*{Create functions to automate your plotting}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Write functions for day to day plots}
<<custom_plots, eval=FALSE, tidy=FALSE>>=
my_custom_plot <- function(df, title = "", ...) {
    ggplot(df, ...) +
    ggtitle(title) +
    whatever geoms() +
    theme(...)
}
@

Then just call your function to generate a plot.
It's a lot easier to fix one function that do it over and over for many plots
<<custom_plots2, eval=FALSE, tidy=FALSE>>=
plot1 <- my_custom_plot(dataset1, title = "Figure 1")
@


\end{frame}

% --------------------------------------------------------------
\section*{Scales}
\frame{\sectionpage}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Commonly used scales}
<<scale_list, eval=FALSE, tidy=FALSE>>=
scale_fill_discrete(), scale_colour_discrete()
scale_fill_hue(), scale_color_hue()
scale_fill_manual(),  scale_color_manual()
scale_fill_brewer(), scale_color_brewer()
scale_linetype(), scale_shape_manual()
@
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Adding a continuous scale}
<<boxplots3 , fig.width=6, fig.height=4, out.width='.75\\linewidth', fig.show='hold',  tidy=FALSE>>=
library(MASS)
ggplot(birthwt, aes(factor(race), bwt)) +
geom_boxplot(width = .2) +
scale_y_continuous(labels = (paste0(1:4, " Kg")),
breaks = seq(1000, 4000, by = 1000))
@
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Another continuous scale with custom labels}
<<scale_1, eval = FALSE, out.width='.75\\linewidth', tidy = FALSE>>=
# Assign the plot to an object
dd <- ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) +
geom_point(size = 4, shape = 16) +
facet_grid(. ~Species)
# Now add a scale
dd +
scale_y_continuous(breaks = seq(2, 8, by = 1),
labels = paste0(2:8, " cm"))
@
\end{frame}


% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{gradients}
<<scale_2, eval = TRUE, fig.width=6, fig.height=4, out.width='.75\\linewidth', tidy = FALSE>>=
h + geom_histogram( aes(fill = ..count..), color="black") +
scale_fill_gradient(low="green", high="red")
@
\end{frame}



% --------------------------------------------------------------
\section*{Publication quality figures}
\frame{\sectionpage}

% How to save your plots
% --------------------------------------------------------------
\begin{frame}[fragile]
\begin{itemize}
\item If the plot is on your screen
<<pub0, eval = FALSE, out.width='.75\\linewidth'>>=
ggsave('~/path/to/figure/filename.png')
@
\item If your plot is assigned to an object
<<pub1, eval = FALSE, out.width='.75\\linewidth'>>=
ggsave(plot1, file = "~/path/to/figure/filename.png")
@

\item Specify a size
<<pub2, eval = FALSE, out.width='.75\\linewidth', tidy = FALSE>>=
ggsave(file = "/path/to/figure/filename.png", width = 6,
height =4)
@
\item or any format (pdf, png, eps, svg, jpg)
<<pub3, eval = FALSE, out.width=".75\\linesewidth">>=
ggsave(file = "/path/to/figure/filename.eps")
ggsave(file = "/path/to/figure/filename.jpg")
ggsave(file = "/path/to/figure/filename.pdf")
@
\end{itemize}
\end{frame}

% --------------------------------------------------------------
\begin{frame}[fragile]
\frametitle{Further help}
\begin{itemize}
\item You've just scratched the surface with ggplot2.
\item Practice
\item Read the docs (either locally in \texttt{R} or at \url{http://docs.ggplot2.org/current/})
\item Work together
\end{itemize}
\includegraphics[scale=.15]{images/chang_book.png}
\includegraphics[scale=.15]{images/hadley.png}
\end{frame}

% --------------------------------------------------------------
% end, hope it was useful.
\end{document}