housing <- read.csv("dataSets/landdata-states.csv")
head(housing[1:5])
housing$Year <- as.numeric(substr(housing$Date, 1, 4))
housing$Qrtr <- as.numeric(substr(housing$Date, 5, 5))
housing$Date <- housing$Year + housing$Qrtr/4
hist(housing$Home.Value)
library(ggplot2)
ggplot(housing, aes(x = Home.Value)) + geom_histogram()
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
plot(Home.Value ~ Date, data=subset(housing, State == "MA"))
points(Home.Value ~ Date, col="red", data=subset(housing, State == "TX"))
legend(19750, 400000, c("MA", "TX"), title="State", col=c("black", "red"), pch=c(1, 1))
ggplot(subset(housing, State %in% c("MA", "TX")), aes(x=Date, y=Home.Value, color=State)) + geom_point()
In ggplot land aesthetic
means “something you can see”. Examples include:
Geometric objects are the actual marks we put on a plot. Examples include:
A plot must have at least one geom; there is no upper limit.
You can add a geom to a plot using the +
operator. You can get a list of available geometric objects using the code: help.search("geom_", package = "ggplot2")
or simply type geom_<tab>
in any good R IDE (such as Rstudio or ESS) to see a list of functions starting with geom_
.
hp2001Q1 <- subset(housing, Date == 2001.25)
ggplot(hp2001Q1,
aes(y = Structure.Cost, x = Land.Value)) +
geom_point()
ggplot(hp2001Q1,
aes(y = Structure.Cost, x = log(Land.Value))) +
geom_point()
A plot constructed with ggplot can have more than one geom. In that case the mappings established in the ggplot() call are plot defaults that can be added to or overridden. Our plot could use a regression line:
hp2001Q1$pred.SC <- predict(lm(Structure.Cost ~ log(Land.Value), data = hp2001Q1))
p1 <- ggplot(hp2001Q1, aes(x = log(Land.Value), y = Structure.Cost))
p1 + geom_point(aes(color = Home.Value)) + geom_line(aes(y = pred.SC))
Not all geometric objects are simple shapes. The smooth geom includes a line and a ribbon.
p1 + geom_point(aes(color = Home.Value)) + geom_smooth()
## `geom_smooth()` using method = 'loess'
Each geom accepts a particualar set of mappings. For example geom_text() accepts a labels mapping.
p1 + geom_text(aes(label=State), size = 3)
## install.packages("ggrepel")
require("ggrepel")
## Loading required package: ggrepel
p1 + geom_point() +geom_text_repel(aes(label=State), size = 3)
Note that variables are mapped to aesthetics with the aes()
function, while fixed aesthetics are set outside the aes() call. This sometimes leads to confusion, as in this example:
p1 +
geom_point(aes(size = 2),# incorrect! 2 is not a variable
color="red") # this is fine -- all points red
Other aesthetics are mapped in the same way as x and y in the previous example.
p1 + geom_point(aes(color=Home.Value, shape = region))
## Warning: Removed 1 rows containing missing values (geom_point).
These data consist of Human Development Index and Corruption Perception Index scores for several countries.
dat <- read.csv("dataSets/EconomistData.csv")
head(dat)
ggplot(dat, aes(x = CPI, y = HDI)) + geom_point(color="blue")
ggplot(dat, aes(x = CPI, y = HDI, color = Region)) + geom_point()
ggplot(dat, aes(x = CPI, y = HDI, color = Region)) + geom_point(size=2)
ggplot(dat, aes(x = CPI, y = HDI)) + geom_point(aes(color = Region, size = HDI.Rank))
Some plot types (such as scatterplots) do not require transformations–each point is plotted at x and y coordinates equal to the original value. Other plots, such as boxplots, histograms, prediction lines etc. require statistical transformations:
Each geom has a default statistic, but these can be changed. For example, the default statistic for geom_bar
is stat_count
:
library(ggplot2)
args(geom_histogram)
## function (mapping = NULL, data = NULL, stat = "bin", position = "stack",
## ..., binwidth = NULL, bins = NULL, na.rm = FALSE, show.legend = NA,
## inherit.aes = TRUE)
## NULL
args(stat_bin)
## function (mapping = NULL, data = NULL, geom = "bar", position = "stack",
## ..., binwidth = NULL, bins = NULL, center = NULL, boundary = NULL,
## breaks = NULL, closed = c("right", "left"), pad = FALSE,
## na.rm = FALSE, show.legend = NA, inherit.aes = TRUE)
## NULL
Arguments to stat_
functions can be passed through geom_
functions. This can be slightly annoying because in order to change it you have to first determine which stat the geom uses, then determine the arguments to that stat.
For example, here is the default histogram of Home.Value:
p2 <- ggplot(housing, aes(x = Home.Value))
p2 + geom_histogram()
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
The binwidth looks reasonable by default, but we can change it by passing the binwidth argument to the stat_bin function:
p2 + geom_histogram(stat = "bin", binwidth=4000)
Sometimes the default statistical transformation is not what you need. This is often the case with pre-summarized data:
housing.sum <- aggregate(housing["Home.Value"], housing["State"], FUN=mean)
rbind(head(housing.sum), tail(housing.sum))
# ggplot(housing.sum, aes(x=State, y=Home.Value)) + geom_bar() # error
What is the problem with the previous plot? Basically we take binned and summarized data and ask ggplot to bin and summarize it again (remember, geom_bar defaults to stat = stat_count); obviously this will not work. We can fix it by telling geom_bar to use a different statistical transformation function:
ggplot(housing.sum, aes(x=State, y=Home.Value)) + geom_bar(stat="identity")
?stat_smooth
.?loess
.ggplot(dat, aes(x=CPI, y=HDI)) + geom_point()
ggplot(dat, aes(x=CPI, y=HDI)) + geom_point() + geom_smooth()
## `geom_smooth()` using method = 'loess'
ggplot(dat, aes(x=CPI, y=HDI)) + geom_point() + geom_smooth(method="lm")
ggplot(dat, aes(x = CPI, y = HDI)) + geom_point() + geom_line(stat = "smooth", method = "loess")
ggplot(dat, aes(x = CPI, y = HDI)) + geom_point() + geom_smooth(span=.4)
## `geom_smooth()` using method = 'loess'
Aesthetic mapping (i.e., with aes()) only says that a variable should be mapped to an aesthetic. It doesn’t say how that should happen. For example, when mapping a variable to shape with aes(shape = x) you don’t say what shapes should be used. Similarly, aes(color = z) doesn’t say what colors should be used. Describing what colors/shapes/sizes etc. to use is done by modifying the corresponding scale. In ggplot2 scales include:
Scales are modified with a series of functions using a **scale_scale_<tab>
to see a list of scale modification functions.
Start by constructing a dotplot showing the distribution of home values by Date and State.
p3 <- ggplot(housing,
aes(x = State, y = Home.Price.Index)) +
theme(legend.position="top", axis.text=element_text(size = 6))
(p4 <- p3 + geom_point(aes(color = Date),
alpha = 0.5,
size = 1.5,
position = position_jitter(width = 0.25, height = 0)))
Now modify the breaks for the x axis and color scales
p4 + scale_x_discrete(name="State Abbreviation") +
scale_color_continuous(name="",
breaks = c(1976, 1994, 2013),
labels = c("'76", "'94", "'13"))
Next change the low and high values to blue and red:
p4 + scale_x_discrete(name="State Abbreviation") +
scale_color_continuous(name="",
breaks = c(1976, 1994, 2013),
labels = c("'76", "'94", "'13"),
low = "blue", high = "red")
library(scales)
p4 + scale_color_continuous(name="",
breaks = c(1976, 1994, 2013),
labels = c("'76", "'94", "'13"),
low = muted("blue"), high = muted("red"))
ggplot2 has a wide variety of color scales; here is an example using scale_color_gradient2 to interpolate between three different colors.
p4 + scale_color_gradient2(name="",
breaks = c(1976, 1994, 2013),
labels = c("'76", "'94", "'13"),
low = muted("blue"),
high = muted("red"),
mid = "gray60",
midpoint = 1994)
ggplot(dat, aes(x=CPI,y=HDI,color=Region)) + geom_point()
ggplot(dat, aes(x=CPI,y=HDI,color=Region)) + geom_point() +
scale_x_continuous(name = "Corruption Perception Index") +
scale_y_continuous(name = "Human Development Index") +
scale_color_discrete(name = "Region of the world")
ggplot(dat, aes(x = CPI, y = HDI, color = Region)) + geom_point() +
scale_x_continuous(name = "Corruption Perception Index") +
scale_y_continuous(name = "Human Development Index") +
scale_color_manual(name = "Region of the world",
values = c("#24576D","#099DD7","#28AADC","#248E84","#F2583F","#96503F"))
Faceting is ggplot2 parlance for small multiples. The idea is to create separate graphs for subsets of data. ggplot2 offers two functions for creating small multiples:
facet_wrap()
: define subsets as the levels of a single grouping variablefacet_grid()
: define subsets as the crossing of two grouping variables Facilitates comparison among plots, not just of geoms within a plotStart by using a technique we already know–map State to color:
p5 <- ggplot(housing, aes(x = Date, y = Home.Value))
p5 + geom_line(aes(color = State))
There are two problems here–there are too many states to distinguish each one by color, and the lines obscure one another.
We can remedy the deficiencies of the previous plot by faceting by state rather than mapping state to color.
(p5 <- p5 + geom_line() + facet_wrap(~State, ncol = 10))
There is also a facet_grid() function for faceting in two dimensions.
The ggplot2 theme system handles non-data plot elements such as
Built-in themes include:
p5 + theme_linedraw()
Specific theme elements can be overridden using theme(). For example:
p5 + theme_minimal() +
theme(text = element_text(color = "red"))
You can create new themes, as in the following example:
theme_new <- theme_bw() +
theme(plot.background = element_rect(size = 1, color = "blue", fill = "black"),
text=element_text(size = 12, color = "ivory"),
axis.text.y = element_text(colour = "purple"),
axis.text.x = element_text(colour = "red"),
panel.background = element_rect(fill = "pink"),
strip.background = element_rect(fill = muted("orange")))
p5 + theme_new
The most frequently asked question goes something like this: I have two variables in my data.frame, and I’d like to plot them as separate points, with different color depending on which variable it is. How do I do that?
housing.byyear <- aggregate(cbind(Home.Value, Land.Value) ~ Date, data = housing, mean)
ggplot(housing.byyear, aes(x=Date)) +
geom_line(aes(y=Home.Value), color="red") +
geom_line(aes(y=Land.Value), color="blue")
library(tidyr)
home.land.byyear <- gather(housing.byyear,
value = "value",
key = "type",
Home.Value, Land.Value)
ggplot(home.land.byyear, aes(x=Date, y=value, color=type)) +
geom_line()
Lets start by creating the basic scatter plot, then we can make a list of things that need to be added or changed. The basic plot looks like this:
dat <- read.csv("dataSets/EconomistData.csv")
pc1 <- ggplot(dat, aes(x = CPI, y = HDI, color = Region))
pc1 + geom_point()
To complete this graph we need to:
Adding the trend line is not too difficult, though we need to guess at the model being displyed on the graph. A little bit of trial and error leds to
(pc2 <- pc1 +
geom_smooth(aes(group = 1),
method = "lm",
formula = y ~ log(x),
se = FALSE,
color = "red")) +
geom_point()
Notice that we put the geom_line layer first so that it will be plotted underneath the points, as was done on the original graph.
This one is a little tricky. We know that we can change the shape with the shape argument, what value do we set shape to? The example shown in ?shape can help us:
## A look at all 25 symbols
df2 <- data.frame(x = 1:5 , y = 1:25, z = 1:25)
s <- ggplot(df2, aes(x = x, y = y))
s + geom_point(aes(shape = z), size = 4) + scale_shape_identity()
## While all symbols have a foreground colour, symbols 19-25 also take a background colour (fill)
s + geom_point(aes(shape = z), size = 4, colour = "Red") +
scale_shape_identity()
s + geom_point(aes(shape = z), size = 4, colour = "Red", fill = "Black") +
scale_shape_identity()
This shows us that shape 1 is an open circle, so
pc2 +
geom_point(shape = 1, size = 4)
That is better, but unfortunately the size of the line around the points is much narrower than on the original.
(pc3 <- pc2 + geom_point(shape = 1, size = 2.5, stroke = 1.25))
This one is tricky in a couple of ways. First, there is no attribute in the data that separates points that should be labelled from points that should not be. So the first step is to identify those points.
pointsToLabel <- c("Russia", "Venezuela", "Iraq", "Myanmar", "Sudan",
"Afghanistan", "Congo", "Greece", "Argentina", "Brazil",
"India", "Italy", "China", "South Africa", "Spane",
"Botswana", "Cape Verde", "Bhutan", "Rwanda", "France",
"United States", "Germany", "Britain", "Barbados", "Norway", "Japan",
"New Zealand", "Singapore")
(pc4 <- pc3 +
geom_text(aes(label = Country),
color = "gray20",
data = subset(dat, Country %in% pointsToLabel)))
This more or less gets the information across, but the labels overlap in a most unpleasing fashion. We can use the ggrepel package to make things better, but if you want perfection you will probably have to do some hand-adjustment.
library("ggrepel")
(pc4 <- pc3 +
geom_text_repel(aes(label = Country),
color = "gray20",
data = subset(dat, Country %in% pointsToLabel),
force = 10))
Things are starting to come together. There are just a couple more things we need to add, and then all that will be left are themeing changes.
Comparing our graph to the original we notice that the labels and order of the Regions in the color legend differ. To correct this we need to change both the labels and order of the Region variable. We can do this with the factor function.
dat$Region <- factor(dat$Region,
levels = c("EU W. Europe",
"Americas",
"Asia Pacific",
"East EU Cemt Asia",
"MENA",
"SSA"),
labels = c("OECD",
"Americas",
"Asia &\nOceania",
"Central &\nEastern Europe",
"Middle East &\nnorth Africa",
"Sub-Saharan\nAfrica"))
pc4$data <- dat
pc4
The next step is to add the title and format the axes. We do that using the scales system in ggplot2.
library(grid)
(pc5 <- pc4 +
scale_x_continuous(name = "Corruption Perceptions Index, 2011 (10=least corrupt)",
limits = c(.9, 10.5),
breaks = 1:10) +
scale_y_continuous(name = "Human Development Index, 2011 (1=Best)",
limits = c(0.2, 1.0),
breaks = seq(0.2, 1.0, by = 0.1)) +
scale_color_manual(name = "",
values = c("#24576D",
"#099DD7",
"#28AADC",
"#248E84",
"#F2583F",
"#96503F")) +
ggtitle("Corruption and Human development"))
Our graph is almost there. To finish up, we need to adjust some of the theme elements, and label the axes and legends. This part usually involves some trial and error as you figure out where things need to be positioned. To see what these various theme settings do you can change them and observe the results.
library(grid) # for the 'unit' function
(pc6 <- pc5 +
theme_minimal() + # start with a minimal theme and add what we need
theme(text = element_text(color = "gray20"),
legend.position = c("top"), # position the legend in the upper left
legend.direction = "horizontal",
legend.justification = 0.1, # anchor point for legend.position.
legend.text = element_text(size = 11, color = "gray10"),
axis.text = element_text(face = "italic"),
axis.title.x = element_text(vjust = -1), # move title away from axis
axis.title.y = element_text(vjust = 2), # move away for axis
axis.ticks.y = element_blank(), # element_blank() is how we remove elements
axis.line = element_line(color = "gray40", size = 0.5),
axis.line.y = element_blank(),
panel.grid.major = element_line(color = "gray50", size = 0.5),
panel.grid.major.x = element_blank()
))
The last bit of information that we want to have on the graph is the variance explained by the model represented by the trend line. Lets fit that model and pull out the R2 first, then think about how to get it onto the graph.
(mR2 <- summary(lm(HDI ~ log(CPI), data = dat))$r.squared)
## [1] 0.5212859
OK, now that we’ve calculated the values, let’s think about how to get them on the graph. ggplot2 has an annotate function, but this is not convenient for adding elements outside the plot area. The grid package has nice functions for doing this, so we’ll use those.
library(grid)
png(file = "econScatter10.png", width = 800, height = 600)
pc6
grid.text("Sources: Transparency International; UN Human Development Report",
x = .02, y = .03,
just = "left",
draw = TRUE)
grid.segments(x0 = 0.81, x1 = 0.825,
y0 = 0.90, y1 = 0.90,
gp = gpar(col = "red"),
draw = TRUE)
grid.text(paste0("R² = ",
as.integer(mR2*100),
"%"),
x = 0.835, y = 0.90,
gp = gpar(col = "gray20"),
draw = TRUE,
just = "left")
dev.off()
## png
## 2