8 Charts

There are various ways to create charts, static or interactive, in R.

8.0.1 Static charts with ggplot2

Static charts can be created with ggplot2 (in PNG format). By default they have a width=7 and height=5. You can alter this in the code chunk options: {r fig.width=8, fig.height=4}
For full reference see: http://ggplot2.tidyverse.org/reference/index.html

library(ggplot2)

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
  geom_point(colour="red", size=2)

# The 1st argument in ggplot() should be the name of your dataframe;
# The 2nd argument should be an aesthetic attribute - this defines which variables will be on the x and y axes.
# You can now add 'Layers' to the ggplot object using the + symbol. You must have at least one 'geom' layer that specifies what type of chart it is (eg. geom_bar, geom_point, geom_line etc..)  

If you only want to chart frequencies, then you only need to specify the x axis variable. In the geom layer add stat=“count” to automatically generate a count.f

ggplot2::ggplot(mtcars, aes(x = hp)) +
 geom_point(stat="count", colour="red") 

8.0.2 Changing Axes

By default the axes labels are based on variable names. To change these or add a title use a ‘labs’ layer:

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
 geom_point(stat="identity", colour="red") +
 labs(title = "Cars", x = 'Horsepower', y = "Miles/Gallon") 

The width of the x and y axes are set automatically, but can be changed:

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
 geom_point(colour="red") +
 labs(title = "Cars", x = 'Horsepower', y = "Miles/Gallon") +
 expand_limits(y = c(0,40), x = c(0,350)) +  #  Customise where the x and y axes start/end
 scale_y_continuous(expand=c(0,0)) +         #  Removes The gap at the bottom of the y axis
 scale_x_continuous(expand=c(0,0))           #  Removes The gap at left of the x axis

8.0.3 Chart Themes & Styles

You can change the appearance of most elements of the chart use a ‘theme’ layer. There are various templates available such as theme_classic (see https://ggplot2.tidyverse.org/reference/ggtheme.html):

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
 geom_point(stat="identity", colour="red") +
 labs(title = "Car Efficiency", x = 'Horsepower', y = "Miles/Gallon") +
 expand_limits(y = 0) +
 scale_y_continuous(expand=c(0,0)) +
 scale_x_continuous(expand=c(0,0)) +
 theme_classic()

You can create a customised theme to : * Change the appeareance of axes (axis.line)
* Change the plot area (panel.background)
* Add/Remove/Change the gridlines (panel.grid.major.x & panel.grid.major.y)
* Change the position and size of the title (plot.title)

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
 geom_point(stat="identity", colour="red") +
 theme(axis.line = element_line(colour = "black",  size = 0.25, linetype = "solid")
      ,axis.ticks = element_blank()
      ,panel.background = element_blank()
      ,panel.grid.major.y = element_line(colour = "blue",  size = 0.1, linetype = "dashed")
      ,panel.grid.major.x = element_line(colour = "green",  size = 0.1, linetype = "dashed")
      ,plot.title = element_text(hjust = 0.5) ) 

You can create reusable styles and themes, which can be applied to all charts:

custom_axis <- list(expand_limits(y = 0, x = 0),
                    scale_y_continuous(expand=c(0,0)), 
                    scale_x_continuous(expand=c(0,0)) )

custom_theme <- theme(axis.line = element_line(colour = "grey",  size = 0.25, linetype = "solid")
                     ,axis.ticks = element_blank()
                     ,panel.background = element_blank()
                     ,panel.grid.major.y = element_line(colour = "grey",  size = 0.1, linetype = "dashed")
                     ,panel.grid.major.x = element_line(colour = "grey",  size = 0.1, linetype = "dashed")
                     ,plot.title = element_text(hjust = 0.5) )

ggplot2::ggplot(mtcars, aes(x = hp, y = mpg)) +
 geom_point(stat="identity", colour="red") +
 custom_axis +   
 custom_theme 

8.0.4 Line Charts (geom_line)

ggplot2::ggplot(mtcars, aes(x = cyl)) +
 geom_line(stat="count", colour="blue", size=2)

8.0.5 Bar Charts (geom_bar & geom_histogram)

geom_bar produces basic bar charts and is best used with categorical variables.
geom_histogram allows you to choose bin widths to automatically re-group you data, so it is better for continuous variables.

ggplot2::ggplot(mtcars, aes(x = hp)) +
 geom_bar(stat="count", fill="orange", width=1)  # Use **colour=** to set the border colour

ggplot2::ggplot(mtcars, aes(x = hp)) + 
  geom_histogram(binwidth = 25, fill="orange")

To flip a bar chart from vertical to horizontal use ‘coord_flip’

ggplot2::ggplot(mtcars, aes(x = hp)) + 
  geom_histogram(binwidth = 25, fill="orange") +
  coord_flip()

8.0.6 Adding features

Extra layers can be added to the chart to show additional information.
Vertical or horizontal lines can be used to show statistics.

ggplot2::ggplot(mtcars, aes(x = hp)) + 
  geom_histogram(binwidth = 25, alpha=0.5, fill="orange") +
  geom_vline(aes(xintercept = median(hp)),col='red',size=1) +
  geom_vline(aes(xintercept = quantile(hp, 0.25)),col='red',size=0.5) +
  geom_vline(aes(xintercept = quantile(hp, 0.75)),col='red',size=0.5)

8.0.7 Heatmaps

esoph %>% group_by(agegp, alcgp) %>% summarise(n = sum(ncontrols)) %>%
          ggplot(aes(x = agegp, y = alcgp, fill = n)) +
           geom_tile()

8.0.8 Boxplots

ggplot2::ggplot(data = mtcars, aes(y = mpg)) +
  geom_boxplot() +
  facet_wrap(vars(cyl))

8.0.9 Combination Charts

You can have multiple layers, for example bars and points

ggplot2::ggplot(mtcars, aes(x = cyl, y = mpg)) +
 geom_bar(stat="identity") +
 geom_point()

8.0.10 Scales

scales package

8.0.11 ggplot wizard

The esquisse package has a shiny wizard that can generate code for different chart types:

esquisse::esquisser()

8.0.12 Interactive charts with ggiraph

The ggiraph package extends ggplot and allows interactivity.

library(ggiraph)

## Warning: package 'ggiraph' was built under R version 3.4.4

8.0.13 Interactive charts with dygraph

For interactive time series charts you can use the dygraph package which is related to the dygraph javascript library.

library(dygraphs)
lungDeaths <- cbind(mdeaths, fdeaths) #Creates the data set for the graph
dygraphs::dygraph(lungDeaths) # Displays the chart with default settings

#It is possible to add a range selector under the chart
dygraphs::dygraph(lungDeaths) %>% dygraphs::dyRangeSelector()

There are lots of options that can be changed such as axis labels, legends, and colours.

#Adding a title 
dygraphs::dygraph(lungDeaths,main='This is the chart title')

#Adding a legend
dygraphs::dygraph(lungDeaths,main='This is the chart title') %>% dygraphs::dyLegend(show='always',width = 500)

#Adding othe options
dygraphs::dygraph(lungDeaths,main='This is the chart title') %>% dygraphs::dyOptions(includeZero = TRUE)