Code Along: Data Visualization with ggplot2

Hosted by the Graduate Students of Open Coding Hour

This 1 hour webinar will focus on making three kinds of visualizations: time-series line graphs, bar charts, and distribution plots. For examples of other kinds of plots (in addition to data transformation vignettes), check the class website for the 4 day R course taught by myself and Cory DuPai: https://rachaelcox.github.io/classes/IntroR_summer_2020.html

September 30, 2020

1. Plotting x-y relationships (e.g. time series, correlations)

We will use the brewmats dataset, which is sourced from the US Alcohol and Tobacco Tax and Trade Bureau (TBB, https://www.ttb.gov/beer/statistics), scraped by the TidyTuesday group and cleaned by me. The brewmats dataset contains the following variables:

  • data_type = denotes that all amounts of material used are given in pounds (lbs)
  • material_type = materials belong in one of two categories, grain (e.g., wheat) or non-grain (e.g., sugar)
  • type = specific material details (e.g., malt, corn, rice, barley, wheat, hops, sugar, etc
  • year = year the amount of materials used was recorded (2008-2015)
  • month = month the amount of materials used was recorded (1=January, 2=February, 3=March, etc)
  • month_usage_by_type = total amount of material usage by type per month (lbs)
  • month_sum_all_types = total amount of material used for all types (lbs)
  • year_sum_by_type = total amount of material used by type per year
Beer Components 
# download the `brewmats` dataset
brewmats <- read_csv("https://rachaelcox.github.io/classes/datasets/brewmats.csv")
## Parsed with column specification:
## cols(
##   data_type = col_character(),
##   material_type = col_character(),
##   year = col_double(),
##   month = col_double(),
##   type = col_character(),
##   month_usage_by_type = col_double(),
##   month_sum_all_types = col_double(),
##   year_sum_by_type = col_double()
## )

Code Along: Plot the total monthly usage of all brewing materials (month_sum_all_types) on the y-axis for every month in the dataset on the x-axis, as a line graph colored by year using geom_line().

# plot the variables as specified
ggplot(brewmats, aes(x = month, y = month_sum_all_types,  # aes maps the variables
                     group = year, color = year)) +
  geom_line()  # calls the line graph

# ggplot is interpreting the numerical variables as continuous
# we use as.factor() to tell R that these variables are discrete
ggplot(brewmats, aes(x = as.factor(month), y = month_sum_all_types,
                     group = as.factor(year), color = as.factor(year))) +
  geom_line() +
  scale_color_colorblind()  # use a colorblind-friendly palette

# clean it up and make it pretty
ggplot(brewmats, aes(x = as.factor(month), y = month_sum_all_types,
                     group = as.factor(year), color = as.factor(year))) +
  geom_line(size = 1.5) +
  scale_color_colorblind(name = "Year") +   # change legend title
  xlab("Month") +   # rename x-axis
  ylab("Monthly Material Usage (lbs)")   # rename y-axis

Practice: Plot the yearly usage (year_sum_by_type) of each type of brewing materials on the y-axis, for every year in the dataset on the x-axis, as a line graph colored by type using geom_line(). Remember to map group = type so that ggplot knows how which lines you want to connect.

ggplot(brewmats, aes(x = as.factor(year), y = year_sum_by_type,
                     group = type, color = type)) +
  geom_line(size = 1.5) +
  scale_color_colorblind(name = "Material Type") +   # change legend title
  scale_y_log10() +  # scale y-axis logarithmically so we can more easily visualize amounts on the lower end
  xlab("Year") +  # rename x-axis
  ylab("Usage (lbs)")  # rename y-axis

2. Plotting amounts (e.g. bar graphs, heat maps)

For this section, we will use the mushrooms dataset (obtained from Kaggle and cleaned by me), which contains the following information:

  • class = whether the mushroom is edible or poisonous
  • cap_shape = shape of the mushroom cap, e.g., bell, conical, convex, flat, etc
  • cap_color = color of the mushroom cap, e.g., brown, buff, cinnamon, gray, etc
  • odor = smell of the mushroom (almond, anise, creosote, fishy foul, musty, pungent, none)
  • gill_spacing = spacing between mushroom gills, aka the underside of the mushroom cap (close, crowded or distant)
  • gill_size = size of mushroom gills (broad or narrow)
  • gill_color = color of mushroom gills (black, brown, etc)
  • stalk_shape = shape of the mushroom stalk (enlarging or tapering)
  • stalk_root = type of stalk root (bulbous, club, cup, rooted, etc)
  • veil_type = type of veil on the mushroom (partial or universal)
  • veil_color = color of the veil (brown, orange, white or yellow)
  • ring_number = number of rings on the stalk of the mushroom (0, 1 or 2)
  • ring_type = description of the ring(s), if any, found on the stalk, e.g., flaring, large, pendant, etc
  • spore_print_color = color of spores collected on a sheet of paper as a print (black, brown, yellow, etc)
  • population = description of nearby mushrooms of the same species, if any; can be abundant, clustered, numerous, scattered, several, and solitary
  • habitat = where the mushroom was found (grasses, leaves, meadows, paths, urban, waste, woods)
Mushroom Diagram 
# download the `mushrooms` dataset
mushrooms <- read_csv("https://rachaelcox.github.io/classes/datasets/mushrooms.csv")
## Parsed with column specification:
## cols(
##   class = col_character(),
##   cap_shape = col_character(),
##   cap_surface = col_character(),
##   cap_color = col_character(),
##   odor = col_character(),
##   gill_spacing = col_character(),
##   gill_size = col_character(),
##   gill_color = col_character(),
##   stalk_shape = col_character(),
##   stalk_root = col_character(),
##   veil_type = col_character(),
##   veil_color = col_character(),
##   ring_number = col_double(),
##   ring_type = col_character(),
##   spore_print_color = col_character(),
##   population = col_character(),
##   habitat = col_character()
## )

Code Along: Plot a bar graph for counts of mushrooms found in each habitat, colored by class (edible or poisonous) using geom_bar().

# plot the variables as specified
ggplot(mushrooms, aes(x = habitat, fill = class)) +  # note we need to use `fill =` instead of `color =` for this kind of plot
  geom_bar()  # default bar position is "stack"

# change bar positioning and fill colors
ggplot(mushrooms, aes(x = habitat, fill = class)) +
  geom_bar(position = "dodge") +  # change bar positions so they are side-by-side
  scale_fill_brewer()  # change the default colors

# make it pretty
ggplot(mushrooms, aes(x = habitat, fill = class)) +
  geom_bar(position = position_dodge(preserve = "single")) +  # fixing bar width requires this special function
  scale_fill_brewer(palette = "Set1") +  # specifying `palette = "Set1"` sets a qualitative color palette
  coord_flip()  # flips the x- and y-axes for better readability

# add more information
ggplot(mushrooms, aes(x = habitat, fill = class)) +
  geom_bar(position = position_dodge(preserve = "single")) +   
  scale_fill_brewer(palette = "Set1", direction = -1) +  # setting `direction = -1` reverses color assignment
  coord_flip() +
  facet_wrap(~population)  # displays habitat and class for each kind of population

Practice: Plot a bar graph for counts of mushrooms of each type of odor, colored by class (edible or poisonous) using geom_bar().

ggplot(mushrooms, aes(x = odor, fill = class)) +
  geom_bar(position = "fill") +  # we get proportions of each class with `position = fill`
  scale_fill_brewer(palette = "Set2") +  # change the color palette
  coord_flip()

ggplot(mushrooms, aes(x = odor, fill = class)) +
  geom_bar(position = "fill") +
  scale_fill_brewer(palette = "Set3") +  # change the color palette again
  coord_flip() +
  facet_wrap(~cap_shape)

3. Plotting distributions (e.g., box plots, histograms, density plots)

For this section, we will use the wine dataset (obtained from Kaggle and cleaned by me), which contains the following variables:

  • type: whether the wine is red or white
  • quality: median score between 0 and 10 as blindly graded by wine experts
  • quality_grade: quality category given to each rating based on distribution of ratings (low, med, high)
  • alcohol: the percent alcohol content of the wine (% by volume)
  • alcohol_grade: relative amount of alcohol compared to all wines (low, med, high)
  • pH: describes how acidic or basic a wine is on a scale from 0 (very acidic) to 14 (very basic); most wines are between 3-4 on the pH scale
  • acidity_grade: acidity intensity (low, med, higj)
  • fixed_acidity: most acids involved with wine or fixed or nonvolatile/do not evaporate readily (tartaric acid - g / dm^3)
  • volatile_acidity: the amount of acetic acid in wine, which at too high of levels can lead to an unpleasant, vinegar taste (acetic acid - g / dm^3)
  • citric_acid: found in small quantities, citric acid can add freshness and flavor to wines (g / dm^3)
  • residual_sugar: the amount of sugar remaining after fermentation stops; it's rare to find wines with less than 1 gram/liter and wines with greater than 45 grams/liter are considered sweet (g / dm^3)
  • chlorides: the amount of salt in the wine (sodium chloride - g / dm^3)
  • free_sulfur_dioxide: the free form of SO2 exists in equilibrium between molecular SO2 (as a dissolved gas) and bisulfite ion; it prevents microbial growth and the oxidation of wine (mg / dm^3)
  • total_sulfur_dioxide: amount of free and bound forms of S02; in low concentrations, SO2 is mostly undetectable in wine, but at free SO2 concentrations over 50 ppm, SO2 becomes evident in the nose and taste of wine (mg / dm^3)
  • density: degree of consistency measured by mass per unit volume (g / cm^3)
  • sulphates: a wine additive which can contribute to sulfur dioxide gas (S02) levels, wich acts as an antimicrobial and antioxidant (potassium sulphate - g / dm^3)
Typical pH of Wine Types

Typical pH of Wine Types

# download the `wine` dataset
wine <- read_csv("https://rachaelcox.github.io/classes/datasets/wine_features.csv")
## Parsed with column specification:
## cols(
##   type = col_character(),
##   quality = col_double(),
##   quality_grade = col_character(),
##   alcohol = col_double(),
##   alcohol_grade = col_character(),
##   pH = col_double(),
##   acidity_grade = col_character(),
##   fixed_acidity = col_double(),
##   volatile_acidity = col_double(),
##   citric_acid = col_double(),
##   residual_sugar = col_double(),
##   chlorides = col_double(),
##   free_sulfur_dioxide = col_double(),
##   total_sulfur_dioxide = col_double(),
##   density = col_double(),
##   sulphates = col_double()
## )

Code Along: Plot the distribution of pH for each wine type (i.e., red or wine) by mapping pH to the x-axis, coloring by type, and calling geom_density(). Then, use geom_boxplot() to visualize the distribution of pH across quality_grade, again coloring by type.

# plot the distribution of pH for red and white wine
ggplot(wine, aes(x = pH, fill = type)) + # note we need to use `fill =` instead of `color =` for this kind of plot
  geom_density()

# make it pretty
ggplot(wine, aes(x = pH, fill = type)) +
  geom_density(alpha = 0.70) +  # reduce opacity of distributions to 70%
  scale_fill_manual(values = c("#790000", "#9e934d"))  # specify custom colors

# plot the distribution of pH for each quality grade for red and white wine
ggplot(wine, aes(x = quality_grade, y = pH, fill = type)) +
  geom_boxplot() +
  scale_fill_manual(values = c("#790000", "#9e934d"))

# make it pretty by reordering the x-axis
ggplot(wine, aes(x = fct_relevel(quality_grade, c("low", "med", "high")), y = pH, fill = type)) +
  geom_boxplot() +
  scale_fill_manual(values = c("#790000", "#9e934d"))

Practice: Choose a numeric variable you are interested in. Plot its distribution relative to a categorical variable, e.g., type, quality_grade, alcohol_grade or acidity_grade. Use geom_density(), geom_boxplot(), and/or geom_violin().

# visualize alcohol grade as it related to residual sugar for both red and white wine
ggplot(wine, aes(x = alcohol_grade, y = residual_sugar, fill = type)) +
  geom_violin() +  # violin plots are an alternative to boxplots
  scale_fill_manual(values = c("#790000", "#9e934d"))

# make it pretty
ggplot(wine, aes(x = fct_relevel(alcohol_grade, c("low", "med", "high")), y = residual_sugar, fill = type)) +
  geom_violin(draw_quantiles = c(0.25, 0.5, 0.75)) +  # we can label intervals in the distribution with `draw_quantiles`
  scale_y_log10() +  # adjust the y-axis for better readability
  scale_fill_manual(values = c("#790000", "#9e934d"), name = "Wine\nType") +
  ylab("Residual sugar (g/dm^3)") +
  xlab("Alcohol grade")