Rstudio GGPLOT2. Preparations. The first plot: Hello world! W2018 RENR690 Zihaohan Sang

Transcription

1 Rstudio GGPLOT2 Preparations There are several different systems for creating data visualizations in R. We will introduce ggplot2, which is based on Leland Wilkinson s Grammar of Graphics. The learning curve is a bit steep, but ultimately you ll be able to produce complex graphs more quickly and easily. You first need to install the ggplot2 and other packages from CRAN: install.packages( ggplot2, gapminder, devtools ) library(ggplot2) ##this is the ggplot2 library(gapminder) ##the example data is from this package Also, to make graphs with ggplot2, the data must be in a data frame, and in long (as opposed to wide) format. The first plot: Hello world! An initial bit of code, to make a scatterplot by gapminder data set: excerpt of the Gapminder data on life expectancy, GDP per capita, and population by country. require(ggplot2) require(gapminder) str(gapminder) ##check data ## Classes 'tbl_df', 'tbl' and 'data.frame': 1704 obs. of 6 variables: ## $ country : Factor w/ 142 levels "Afghanistan",..: ## $ continent: Factor w/ 5 levels "Africa","Americas",..: ## $ year : int ## $ lifeexp : num ## $ pop : int ## $ gdppercap: num ggplot(data = gapminder, mapping = aes(x=gdppercap, y=lifeexp)) + geom_point() 1

2 It s not ideal, but not bad as the very first plot. ggplot() function creates a graphics object(like plot() function in base R plotting); then, another two key concepts in the grammar of graphics: aesthetics map features of the data (for example, the lifeexp variable) to features of the visualization (for example, the y-axis coordinate), and geoms concern what actually gets plotted (here, each data point becomes a point in the plot). You could also save the output as one object, like this: a = ggplot(data = gapminder, mapping = aes(x = gdppercap, y = lifeexp)) + geom_point() ##output is saved as a object, and no graph will show ##if you want to check/edit it, just print it out print(a) Why don t we color the points by location (continents)? maybe we could also change the size of points based on the population size? Of course we CAN. ggplot(data = gapminder, mapping = aes(x = gdppercap, y = lifeexp, col = continent, size = pop)) + geom_point() It s best to do the x-axis on a log scale. b <- ggplot(gapminder, aes(x=gdppercap, y=lifeexp, col = continent, size = pop)) + geom_point() + scale_x_log10() print(b) Looks much better! But there are too much points in the plot, why don t we seperate the dataset into several smaller ones(based on certain variable) and plots them one by one. Faceting A particularly valuable feature of ggplot2 is faceting: the ability to make a series of plots, conditional on the values of some selected variables. 2

3 For example, rather than coloring points by continent, one might separate the continents into separate panels. There are two functions for this, facet_grid() and facet_wrap(). The facet_wrap function is particularly useful if you have a lot of facets, for example, plotting by year. b + facet_wrap(~ year) Other geoms We ve focused so far on scatterplots, but one can also create one-dimensional summaries, such as histograms or boxplots. For a histogram, you want only the x aesthetic, and then use geom_histogram(), with binwidth to define the width of the bins. Here s a histogram of lifeexp for gm_2007 <- gapminder[gapminder$year==2007,] ggplot(gm_2007, aes(x=lifeexp)) + geom_histogram(binwidth=2) 3

4 Alternatively, we can look at boxplots, for which you need to define a continuous variable for y and a categorical variable for x. ggplot(gm_2007, aes(y=lifeexp, x=continent)) + geom_boxplot() Change theme The default of ggplots has gray background. If you don t like it, we can also change it by adding theme_... functions. b + theme_bw() b + theme_classic() b + theme_gray() bonus GGplot2 is fun! Not only can make plots, ggplot2 can also make simple animation. require(gganimate) devtools::install_github( dgrtwo/gganimate ) ## Loading required package: gganimate c = ggplot(gapminder, aes(x=gdppercap, y=lifeexp, col = continent, size = pop, frame = year)) + geom_point() + scale_x_log10() gganimate(c, "output.html") Intermediate ggplot2 Plotting means and error bars The example below is generated from ToothGrowth dataset: the response of the length of odontoblasts (cells responsible for tooth growth) in 60 guinea pigs. Each animal received one of 4

5 three dose levels of vitamin C (0.5, 1, and 2 mg/day) by one of two delivery methods, (orange juice or ascorbic acid (a form of vitamin C and coded as VC). Len represents mean tooth length(mm) of certain treatment, and standard deviation (sd), standard error (se) and 95% confident interval (ci) can be found in example.csv. dat <- read.csv("example2.csv") str(dat) ## 'data.frame': 6 obs. of 7 variables: ## $ supp: Factor w/ 2 levels "OJ","VC": ## $ dose: num ## $ N : int ## $ len : num ## $ sd : num ## $ se : num ## $ ci : num ##once we load the data, we can make the graph ##basic line plot with different color for each supplement type (VC or OJ) ggplot(dat, aes(x=dose, y=len, colour=supp, linetype = supp)) + geom_line() + geom_point() # add errorbar(standard error of the mean) ggplot(dat, aes(x=dose, y=len, colour=supp, linetpe = supp)) + geom_errorbar(aes(ymin=len-se, ymax=len+se), width=.1) + geom_line() + geom_point() # The errorbars overlapped, so use position_dodge to move them horizontally pd <- position_dodge(0.1) # move them.05 to the left and right ggplot(dat, aes(x=dose, y=len, colour=supp, linetype = supp)) + geom_errorbar(aes(ymin=len-se, ymax=len+se), width=.1, position=pd) + geom_line(position=pd) + geom_point(position=pd) 5

6 ggplot(dat, aes(x=dose, y=len, colour=supp, linetype = supp, group=supp)) + geom_errorbar(aes(ymin=len-se, ymax=len+se), colour="black", width=.1, position=pd) + geom_line(position=pd) + geom_point(position=pd, size=3, shape=21, fill="white") + # 21 is filled circle xlab("dose (mg)") + #label on x axis ylab("tooth length(mm)") + #label on y axis scale_colour_hue(name="supplement type", # Legend label, use darker colors breaks=c("oj", "VC"), labels=c("orange juice", "Ascorbic acid"), l=40) + # Use darker colors, lightness=40 ggtitle("the Effect of Vitamin C on\ntooth Growth in Guinea Pigs") + expand_limits(y=0) + # Expand y range scale_y_continuous(breaks=0:20*4) + # Set tick every 4 theme_bw() + theme(legend.justification=c(1,0), legend.position=c(1,0)) # Position legend in bottom right A finished graph with error bars representing the standard error of the mean might look like this. Bar graphs The procedure can also be used to plot bar graphs, but bar plots only work for factors. # Use dose as a factor rather than numeric dat$dose <- factor(dat$dose) str(dat) # Error bars represent standard error of the mean ggplot(dat, aes(x=dose, y=len, fill=supp)) + geom_bar(position=position_dodge(), stat="identity") + geom_errorbar(aes(ymin=len-se, ymax=len+se), width=.2, # Width of the error bars position=position_dodge(.9)) + scale_fill_gray() 6