1 Outline

Import Apple Mobility Data
Wrangle Apple Mobility Data
Counting things
Histograms
Density plots
Ridgeline plots
Violin plots

2 Objectives

In order to use graphs and figures to effectively communicate with our audience, we need to consider a few things:

Recognize the needs of your audience (who are they, and where are they coming from?)

level of data literacy, subject matter expertise, etc.

Communicate the quality of the data with stakeholders (can we answer their question(s) with the available data?)

let them know the good and the bad news

Identify the correct data visualization (based on the data and the problem/question)

single variable, bivariate, and multivariate graphs

3 Materials

View the slides for this section here.

4 More resources

Read more about ggplot2 on the tidyverse website, and in the Data Visualisation chapter of R for Data Science.

5 Load the packages

The main packages we’re going to use are dplyr, tidyr, and ggplot2. These are all part of the tidyverse, so we’ll import this package below:

install.packages("tidyverse")
library(tidyverse)

6 Example: COVID and Mobility

Assume we received the following questions:

How has COVID changed our modes of transportation?

Are people using fewer or different forms of transportation since the COVID pandemic?

Questions we should be considering:

What kind of measurements would this be?

how people travel (walk, drive, etc.)

What would these data look like?

what would the columns and rows look like?

6.1 Data Import

We did some digging and came up with the following dataset to try and answer the questions above:

Apple mobility data: https://covid19.apple.com/mobility

6.1.1 Import

Import the data below

AppleMobRaw <- readr::read_csv("https://bit.ly/3DEDa8T")

View the head() and tail()

head(AppleMobRaw)

tail(AppleMobRaw)

6.1.2 Tidy

We can see the dates are structured across the columns, so we need to restructure these into a tidy format. Read more about this format here.

AppleMobRaw %>% 
  tidyr::pivot_longer(cols = -c(geo_type:country), 
                      names_to = "date", 
                      values_to = "dir_request")

6.1.3 Missing

Now that we have the data in a tidy format, we should remove the missing values from country and sub-region

AppleMobRaw %>% 
  tidyr::pivot_longer(cols = -c(geo_type:country), 
                      names_to = "date", values_to = "dir_request") %>% 
    # remove missing country and missing sub-region data
  dplyr::filter(!is.na(country) & !is.na(`sub-region`))

6.1.4 Mutate

Use mutate() to create a properly formatted date variable, and rename() the transportation_type variable to trans_type. Apply janitor::clean_names() to the entire dataset and assign the final output to TidyApple.

AppleMobRaw %>% 
  tidyr::pivot_longer(cols = -c(geo_type:country), 
                      names_to = "date", values_to = "dir_request") %>% 
    # remove missing country and missing sub-region data
  dplyr::filter(!is.na(country) & !is.na(`sub-region`)) %>% 
  # format date
  mutate(date = lubridate::ymd(date)) %>% 
  # change name of transportation types
  rename(trans_type = transportation_type) %>% 
  # clean names 
  janitor::clean_names() -> TidyApple

6.2 Counting

One of the most important jobs of analytic work is counting things. There are many ways to accomplish this in R, but we’ll stick with the dplyr package because it’s part of the tidyverse.

The dplyr function for counting responses of a categorical or factor variable is count(), and it works like this:

Data %>% 
  count(variable)

6.2.1 Count

So, if we wanted to count the number of different transportation types in the TidyApple data frame, it would look like this,

TidyApple %>% 
  dplyr::count(trans_type)

6.2.2 Sort

We can also sort the responses using the sort = TRUE argument.

TidyApple %>% 
  dplyr::count(trans_type, sort = TRUE)

6.2.3 Iterate

We can also combine dplyr::select_if() and purrr::map() to pass the count() function to all the character variables in TidyApple.

TidyApple %>% 
  select_if(is.character) %>% 
  map(~count(data.frame(x = .x), x, sort = TRUE)) -> tidy_apple_counts

We can example the counts of each value by using the $ to subset the tidy_apple_counts list.

tidy_apple_counts$sub_region

tidy_apple_counts$region

7 Visualizing Distributions

Before we start looking at relationships between variables, we should examine each variable’s underlying distribution. In the next section, we’re going to cover a few graphs that display variable distributions: histograms, density, violin, and ridgeline plots,

7.1 Histograms

A histogram is a special kind of bar graph–it only takes a single continuous variable (in this case, dir_request), and it displays a relative breakdown of the values.

The x axis for the histogram will have the direction requests, and the y variable will display a count of the values.

lab_hist <- labs(x = "Apple directions requests",
                 y = "Count",
     title = "Distribution of Direction Requests",
     subtitle = "source: https://covid19.apple.com/mobility")

7.1.1 exercise

Create a histogram of direction requests using dir_request

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = ____________)) + 
  lab_hist

7.1.2 solution

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request)) + 
  lab_hist

7.2 Adjusting Y Axes

We can see the y axis of the histogram is in scientific notation. This might be hard for some audiences to interpret, so we will change this to use the whole number with commas with the scales package.

7.2.1 exercise

Add the scales::comma value to the scale_y_continuous() function.

library(scales)
TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request)) + 
  scale_y_continuous(labels = __________) +
  lab_hist

7.2.2 solution

library(scales)
TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request)) + 
  scale_y_continuous(labels = scales::comma) +
  lab_hist

7.3 Histogram Shape

We can control the shape of the histogram with the bins argument. The default is 30.

7.3.1 exercie

Set bins to 15.

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request), bins = __) + 
  scale_y_continuous(labels = scales::comma) +
  lab_hist

7.3.2 solution

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request), bins = 15) + 
  scale_y_continuous(labels = scales::comma) +
  lab_hist

7.3.3 exercise

Set bins to 45 and assign it to gg_hist45.

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request), bins = __) + 
  scale_y_continuous(labels = scales::comma) +
  lab_hist -> _____________

7.3.4 solution

TidyApple %>% ggplot() + 
  geom_histogram(aes(x = dir_request), bins = 45) + 
  scale_y_continuous(labels = scales::comma) +
  lab_hist -> gg_hist45

7.4 Density Plots

What if we want to see how a continuous variable is distributed across a categorical variable? We covered this in the previous lesson with a boxplot.

Density plots come in handy here (so do geom_boxplot()s!). Read more about the density geom here.

We are going to create the graph labels so we know what to expect when we build our graph, and we want to see the distribution of the directions request, filled by the levels of transportation type.

lab_density <- labs(x = "Apple directions requests",
                    fill = "Transit Type",
     title = "Distribution of Direction Requests vs. Transportation Type",
     subtitle = "source: https://covid19.apple.com/mobility")

Now we build the density plot, passing the variables so they match our labels above.

7.4.1 exercise

Create a density plot of direction requests colored by the type of transportation.

TidyApple %>% 
  ggplot() +
  geom_density(aes(x = __________, fill = __________)) + 
  lab_density

7.4.2 solution

One drawback to density plots is the y axis can be hard to interpret

TidyApple %>% 
  ggplot() +
  geom_density(aes(x = dir_request, fill = trans_type)) + 
  lab_density

7.4.3 exercise

Adjust the overlapping densities by setting alpha to 1/3. Assign this plot to gg_density.

TidyApple %>% 
  ggplot() +
  geom_density(aes(x = dir_request, fill = trans_type), 
               alpha = __________) + 
  
  lab_density -> __________

7.4.4 solution

TidyApple %>% 
  ggplot() +
  geom_density(aes(x = dir_request, fill = trans_type), 
               alpha = 1/3) + 
  lab_density -> gg_density

gg_density

7.5 Ridgeline Plots

Another option is a ridgeline plot (from the ggridges package). These display multiple densities.

lab_ridges <- labs(
     title = "Direction Requests by Transportation Type",
     subtitle = "source: https://covid19.apple.com/mobility",
     fill = "Transit type",
     x = "Apple directions requests",
     y = "Transportation Types")

library(ggridges)
TidyApple %>%  
  ggplot() + 
  geom_density_ridges(aes(x = dir_request, 
                          y = trans_type, 
                          fill = trans_type), 
                      alpha = 1/5) + 
  lab_ridges

7.6 Violin Plots

Another alternative to the density plot is the violin plot.

7.6.1 exercise

assign "Apple directions requests" to the x axis
assign "Transit Type" to the y axis

lab_violin <- labs(x = _________________________,
                    y = _________________________,
                   fill = "Transit Type",
     title = "Distribution of Direction Requests vs. Transportation Type",
     subtitle = "source: https://covid19.apple.com/mobility")

7.6.2 solution

lab_violin <- labs(x = "Transit Type",
                   y = "Apple directions requests",
                   fill = "Transit Type",
     title = "Distribution of Direction Requests vs. Transportation Type",
     subtitle = "source: https://covid19.apple.com/mobility")

7.6.3 exercise

Add a geom_violin() to the code below:

TidyApple %>% 
  ggplot() +
  ____________(aes(y = dir_request, x = trans_type, 
                  fill = trans_type)) + 
  lab_violin

7.6.4 solution

TidyApple %>% 
  ggplot() +
  geom_violin(aes(y = dir_request, x = trans_type, 
                  fill = trans_type)) + 
  lab_violin

7.6.5 exercise

The great thing about ggplot2s layered syntax, is that we can add geoms with similar aesthetics to the same graph! For example, we can see how geom_violins and geom_boxplots are related by adding a geom_boxplot() layer to the graph above.

TidyApple %>% 
  ggplot() +
  geom_violin(aes(y = dir_request, x = trans_type, 
                  fill = trans_type), alpha = 1/5) + 
  ___________(aes(y = dir_request, x = trans_type, 
                   color = trans_type)) + 
  lab_violin

7.6.6 solution

Note we set the alpha to 1/5 for the geom_violin(), and the color to trans_type for the geom_boxplot().

TidyApple %>% 
  ggplot() +
  geom_violin(aes(y = dir_request, x = trans_type, 
                  fill = trans_type), alpha = 1/5) + 
  geom_boxplot(aes(y = dir_request, x = trans_type, 
                   color = trans_type)) + 
  lab_violin

7.7 Export the data

You’ll want to export the TidyApple dataset for the next set of exercise.

The code chunk below exports the dataset as a .csv.

fs::dir_create("../data/wk5-01-intro-to-ggp2-part-02/")
readr::write_csv(x = TidyApple, 
                 file = "../data/wk5-01-intro-to-ggp2-part-02/TidyApple.csv")

Data Visualization with ggplot2 (part 1)