• pivot_longer() helps us take our data from wide to long format
  • names_to = gives a new name to the pivoted columns
    
  • values_to = gives a new name to the values that used to be in those columns
    
• pivot_wider() helps us take our data from long to wide format
  • names_from specifies the old column name that contains the new column names
    
  • values_from specifies the old column name that contains new cell values
    
• to merge/join data sets together need a variable in common - usually “id”

📃Cheatsheet

• to merge/join data sets together need a variable in common - usually “id”
• ?join - see different types of joining for dplyr
• inner_join(x, y) - only rows that match for x and y are kept
• full_join(x, y) - all rows of x and y are kept
• left_join(x, y) - all rows of x are kept even if not merged with y
• right_join(x, y) - all rows of y are kept even if not merged with x
• anti_join(x, y) - all rows from x not in y keeping just columns from x.
• esquisser() function of the esquisse package can help make plot sketches

📃Cheatsheet

More customization:

• branding
• making plots interactive
• combining plots

Easier plot automation (creating plots in scripts)

Faster (eventually)

• A package for producing graphics - gg = Grammar of Graphics

• Created by Hadley Wickham in 2005

• Belongs to “Tidyverse” family of packages

• “Make a ggplot” = Make a plot with the use of ggplot2 package

Resources:

• https://ggplot2-book.org/

• https://www.opencasestudies.org/

Based on the idea of:

plot objects are placed on top of each other with +

📉 +
📈

Slide Credit: Tanya Shapiro

• Pros: extremely powerful/flexible – allows combining multiple plot elements together, allows high customization of a look, many resources online

• Cons: ggplot2-specific “grammar of graphic” of constructing a plot

• ggplot2 gallery

To make graphics using ggplot2, our data needs to be in a tidy format

Tidy data:

1. Each variable forms a column.
2. Each observation forms a row.

Messy data:

• Column headers are values, not variable names.
• Multiple variables are stored in one column.
• Variables are stored in both rows and columns.

Ideally we want each variable as a column and we want each observation in a row.

Column headers are values, not variable names:

Read more about tidy data and see other examples: Tidy Data tutorial

Let’s plot the CO heat-related ER visits dataset we’ve been working with. First, we’ll only consider data from Boulder county.

Is the data in tidy? Is it in long format?

er <- 
  read_csv("https://daseh.org/data/CO_ER_heat_visits.csv")
er_Boulder <- er %>% filter(county == "Boulder")

head(er_Boulder)

# A tibble: 6 × 6
  county   rate lower95cl upper95cl visits  year
  <chr>   <dbl>     <dbl>     <dbl>  <dbl> <dbl>
1 Boulder  4.03      2.05      6.67     12  2011
2 Boulder  4.08      2.15      6.62     13  2012
3 Boulder  3.79      1.90      6.32     12  2013
4 Boulder  6.29      3.71      9.54     19  2014
5 Boulder  4.76      2.57      7.61     14  2015
6 Boulder  5.68      3.31      8.67     18  2016

Will set up the plot - it will be empty!

• Aesthetic mapping aes(x= , y =) describes how variables in our data are mapped to elements of the plot - Note you don’t need to use mapping but it is helpful to know what we are doing.

library(ggplot2) # don't forget to load ggplot2
# This is not code but shows the general format
ggplot({data_to plot}, aes(x = {var in data to plot},
                                      y = {var in data to plot}))

ggplot(er_Boulder, aes(x = year, y = rate))

There are many to choose from, to list just a few:

• geom_point() – points (we have seen)
• geom_line() – lines to connect observations
• geom_boxplot() – boxplots
• geom_histogram() – histogram
• geom_bar() – bar plot
• geom_col() – column plot
• geom_tile() – blocks filled with color

When to use what plot? A few examples:

• a scatterplot (geom_point()): to examine the relationship between two sets of continuous numeric data
• a barplot (geom_bar()): to compare the distribution of a quantitative variable (numeric) between groups or categories
• a histogram (geom_hist()): to observe the overall distribution of numeric data
• a boxplot (geom_boxplot()): to compare values between different factor levels or categories

Need the + sign to add the next layer to specify the type of plot

ggplot({data_to plot}, aes(x = {var in data to plot},
                                      y = {var in data to plot})) +
  geom_{type of plot}</div>

ggplot(er_Boulder, aes(x = year, y = rate)) +
  geom_point()

Read as: using CO statewide ER heat visits data, and provided aesthetic mapping, add points to the plot

Having the + sign at the beginning of a line will not work!

ggplot(er_Boulder, aes(x = year,
                           y = rate,
                           fill = item_categ))  
 + geom_boxplot()

Pipes will also not work in place of +!

ggplot(er_Boulder, aes(x = year,
                           y = rate,
                           fill = item_categ))  %>%
geom_boxplot()

plt1 <- ggplot(er_Boulder, aes(x = year, y = rate)) +
          geom_point()

plt2 <- ggplot(er_Boulder, aes(x = year, y = rate)) +
          geom_line()

plt1 # fig.show = "hold" makes plots appear
plt2 # next to one another in the chunk settings

Layer a plot on top of another plot with +

ggplot(er_Boulder, aes(x = year, y = rate)) +
  geom_point() +
  geom_line()

Let’s map ER visit rates for four CO counties on the same plot

er_visits_4 <- er %>% 
  filter(county %in% c("Denver", "Weld", "Pueblo", "Jackson"))

ggplot(er_visits_4, aes(x = year, y = rate, color = county)) +
  geom_point() +
  geom_line()

There’s not only the built in ggplot2 themes but all kinds of themes from other packages! - ggthemes - ThemePark package - hrbr themes

You can use theme(legend.position = "none") to remove the legend.

er_visits_4 %>% ggplot(aes(x = year,
                      y = rate,
                      color = county)) +
  geom_line(size = 0.8) +
  theme(legend.position = "none")

It’s possible to go in and change specifications with newer layers. Here is our original plot.

er_visits_4 %>% ggplot(aes(x = year,
                      y = rate,
                      color = county)) +
  geom_line(size = 0.8)

It’s possible to go in and change specifications with newer layers.

er_visits_4 %>% ggplot(aes(x = year,
                      y = rate,
                      color = county)) +
  geom_line(size = 0.8, color = "black")

A. Add a plot_ layer like plot_point()

B. Add a geom_ layer like geom_point()

A. %>%

B. &

C. +

• ggplot() specifies what data use and what variables will be mapped to where
• inside ggplot(), aes(x = , y = , color =) specify what variables correspond to what aspects of the plot in general
• layers of plots can be combined using the + at the end of lines
• special theme_*() functions can change the overall look
• individual layers can be customized using arguments like: size, color alpha (more transparent is closer to 0), and linetype
• labels can be added with the labs() function and x, y, title arguments
• scale_x_continuous() and scale_y_continuous() can modify the scale of the axes
• by default, ggplot() removes points with missing values from plots.

🏠 Class Website

💻 Lab

📃Day 6 Cheatsheet

The theme() function can help you modify various elements of your plot. Here we will adjust the font size of the plot title.

ggplot(er_Boulder, aes(x = year, y = rate)) +
            geom_point(size = 5, color = "green", alpha = 0.5) +
            geom_line(size = 0.8, color = "blue", linetype = 2) + 
  labs(title = "Heat-Related ER Visits:Boulder") +
  theme(plot.title = element_text(size = 20))

The theme() function always takes:

1. an object to change (use ?theme() to see - plot.title, axis.title, axis.ticks etc.)
2. the aspect you are changing about this: element_text(), element_line(), element_rect(), element_blank()
3. what you are changing:
   • text: size, color, fill, face, alpha, angle
   • position: "top", "bottom", "right", "left", "none"
   • rectangle: size, color, fill, linetype
   • line: size, color, linetype

The theme() function can help you modify various elements of your plot. Here we will adjust the horizontal justification (hjust) of the plot title.

ggplot(er_Boulder, aes(x = year, y = rate)) +
            geom_point(size = 5, color = "green", alpha = 0.5) +
            geom_line(size = 0.8, color = "blue", linetype = 2) + 
  labs(title = "Heat-Related ER Visits:Boulder") +

  theme(plot.title = element_text(hjust = 0.5, size = 20))

ggplot(er_Boulder, aes(x = year, y = rate)) +
            geom_point(size = 5, color = "green", alpha = 0.5) +
            geom_line(size = 0.8, color = "blue", linetype = 2) + 
  labs(title = "Heat-Related ER Visits: Boulder") +
  theme(plot.title = element_text(hjust = 0.5, size = 20),
        axis.title = element_text(size = 16))

https://github.com/claragranell/ggplot2/blob/main/ggplot_theme_system_cheatsheet.pdf

Let’s start with er_visits_4.

ggplot(er_visits_4, aes(x = year,
                 y = rate)) +
  geom_line()

You can use group element in a mapping to indicate that each county will have a rate line.

ggplot(er_visits_4, aes(x = year,
                           y = rate,
                           group = county)) +
  geom_line()

ggplot(er_visits_4, aes(x = year,
                           y = rate,
                       color = county)) +
  geom_line()+
  theme(legend.position = "bottom")

Let’s talk additional tricks and tips for making ggplots!

Can add width argument to make the jitter more narrow.

ggplot(er_visits_4, aes(x = county,
                 y = rate,
                 fill = county)) +
  geom_boxplot() +
  geom_jitter(width = .06)

scale_fill_viridis_d() for discrete /categorical data scale_fill_viridis_c() for continuous data

ggplot(er_visits_4, aes(x = county,
                 y = rate,
                 fill = county)) +
  geom_boxplot() +
  geom_jitter(width = .06) +
  scale_fill_viridis_d()

er_bar <- er_visits_4 %>%
  group_by(county) %>%
  summarize("max_rate" = max(rate, na.rm=T)) %>%

ggplot(aes(x = county,
           y = max_rate,
           fill = county)) +
  scale_fill_viridis_d()+
  geom_col() +
  theme(legend.position = "none")

er_bar

Can be used to add an outline around column/bar plots.

er_bar +
   geom_col(color = "black")

er_bar +
  theme(text = element_text(size = 20))

lots_of_lines +
facet_grid( ~ county) +
theme(legend.position = "none") +
theme(axis.text.x = element_text(angle = 90))

#install.packages("plotly")
library(plotly) # creates interactive plots!
ggplotly(lots_of_lines)

Also check out the ggiraph package

Great for combining plots together

Also check out the patchwork package

#install.packages("patchwork")
library(patchwork)
lots_of_lines + er_bar

A few options:

• RStudio > Plots > Export > Save as image / Save as PDF
• RStudio > Plots > Zoom > [right mouse click on the plot] > Save image as
• In the code

ggsave(filename = "saved_plot.png",  # will save in working directory
       plot = rp_fac_plot,
       width = 6, height = 3.5)               # by default in inches

A. Use the fill argument in the aes specification

B. Use color argument in geom_boxplot()

A. add more theme() layers

B. Use facet_grid() to split the plot up

• The theme() function helps you specify aspects about your plot
  • move or remove a legend with theme(legend.position = "none")
  • change font aspects of individual text elements theme(plot.title = element_text(size = 20))
  • center a title: theme(plot.title = element_text(hjust = 0.5))
• sometimes you need to add a group element to aes() if your plot looks strange
• make sure you are plotting what you think you are by checking the numbers!
• facet_grid(~ variable) and facet_wrap(~variable) can be helpful to quickly split up your plot
  • facet_wrap() allows for a scales = "free" argument so that you can have a different axis scale for different plots
• use fill to fill in boxplots

Check out this guide for more information!

🏠 Class Website

💻 Lab

📃Day 6 Cheatsheet

📃Posit’s theme cheatsheet

Image by Gerd Altmann from Pixabay

You can change look of each layer separately. Note the arguments like linetype and alpha that allow us to change the opacity of the points and style of the line respectively.

ggplot(er_Boulder, aes(x = year, y = rate)) +
  geom_point(size = 5, color = "red", alpha = 0.5) +
  geom_line(size = 0.8, color = "black", linetype = 2)

linetype can be given as a number. See the docs for what numbers correspond to what linetype!

You can change the look of whole plot - specific elements, too - like changing font and font size - or even more fonts

ggplot(er_Boulder, aes(x = year, y = rate)) +
  geom_point(size = 5, color = "green", alpha = 0.5) +
  geom_line(size = 0.8, color = "blue", linetype = 2) +
  theme_bw() +
  theme(text=element_text(size=16,  family="Comic Sans MS"))

xlim() and ylim() can specify the limits for each axis

ggplot(er_Boulder, aes(x = year, y = rate)) +
            geom_point(size = 5, color = "green", alpha = 0.5) +
            geom_line(size = 0.8, color = "blue", linetype = 2) +
  labs(title = "Heat-Related ER Visits in Boulder Co.",
              x = "Year",
              y = "Age-adjusted Visit Rate") +
  ylim(0, max(pull(er_visits_4, rate)))

linetype

source

• geoms that draw lines have a linetype parameter
• these include values that are strings like “blank”, “solid”, “dashed”, “dotdash”, “longdash”, and “twodash”

er_Boulder %>% ggplot(aes(x = year,
                      y = rate)) +
  geom_line(size = 0.8, linetype = "twodash")

shape

source

• geoms that draw have points have a shape parameter
• these include numeric values (don’t need quotes for these) and some characters values (need quotes for these)

er_Boulder %>% ggplot(aes(x = year,
                      y = rate)) +
  geom_point(size = 2, shape = 12)

Guide on how to: https://rpubs.com/mclaire19/ggplot2-custom-themes

Set outlier.shape = NA to get ride of outliers. Be careful about if you really should remove these!

However, if can be helpful if your plot is getting stretched to accommodate plotting an outlier. You can always say in the figure legend what you removed.

er_no_out1 <- ggplot(er_visits_4, aes(y = visits, x = county)) +
  geom_boxplot()

er_no_out2 <- ggplot(er_visits_4, aes(y = visits, x = county)) +
  geom_boxplot(outlier.shape = NA) +
  ylim(0,40)

Great for adding labels directly onto plots https://www.opencasestudies.org/ocs-bp-co2-emissions/

#install.packages("directlabels")
library(directlabels)
direct.label(lots_of_lines, method = list("angled.boxes"))