Recreating (more) data visualizations from the book “Knowledge is Beautiful”: Part IV
Welcome to the last part of the series where I recreate data visualizations in R from the book Knowledge is Beautiful by David McCandless.
Links to part I, II, III of the series can be found here.
Plane Crashes
This dataset will be used for a couple of visualizations.
The first visualization is a stacked-barplot showing causes of crashes for every plane crash from 1993 to January 2017 (for flights that were not military, medical or a private chartered flight).
library(dplyr)
library(ggplot2)
library(tidyr)
library(extrafont)df <- read.csv("worst_plane.csv")# Drop the year plane model entered service
mini_df <- df %>%
select(-year_service) %>%
# Gather the wide dataframe into a tidy format
gather(key = cause, value = proportion, -plane)
# Order by cause
mini_df$cause <- factor(mini_df$cause, levels = c("human_error","weather", "mechanical", "unknown", "criminal"), ordered = TRUE)# Create vector of plane names according to year they entered service
names <- unique(mini_df$plane)
names <- as.vector(names)
# sort by factor
mini_df$plane <- factor(mini_df$plane, levels = names)ggplot(mini_df, aes(x=plane, y=proportion, fill=cause)) +
geom_bar(stat = "identity") +
coord_flip() +
# Reverse the order of a categorical axis
scale_x_discrete(limits = rev(levels(mini_df$plane))) +
# Select manual colors that McCandless used
scale_fill_manual(values = c("#8E5A7E", "#A3BEC7", "#E1BD81", "#E9E4E0", "#74756F"), labels = c("Human Error", "Weather", "Mechanical", "Unknown", "Criminal")) +
labs(title = "Worst Planes", caption = "Source: bit.ly/KIB_PlaneCrashes") +
scale_y_reverse() +
theme(legend.position = "right",
panel.background = element_blank(),
plot.title = element_text(size = 13,
family = "Georgia",
face = "bold", lineheight = 1.2),
plot.caption = element_text(size = 5,
hjust = 0.99, family = "Georgia"),
axis.text = element_text(family = "Georgia"),
# Get rid of the x axis text/title
axis.text.x=element_blank(),
axis.title.x=element_blank(),
# and y axis title
axis.title.y=element_blank(),
# and legend title
legend.title = element_blank(),
legend.text = element_text(family = "Georgia"),
axis.ticks = element_blank())
The second visualization is an alluvial diagram for which we can use the ggalluvial package. I should mention that the original visualization by McCandless is much fancier than what this produces but displays the same basic information.
library(alluvial)
library(ggalluvial)crash <- read.csv("crashes_alluvial.csv")# stratum = cause, alluvium = freqggplot(crash, aes(weight = freq,
axis1 = phase,
axis2 = cause,
axis3 = total_crashes)) +
geom_alluvium(aes(fill = cause),
width = 0, knot.pos = 0, reverse = FALSE) +
guides(fill = FALSE) +
geom_stratum(width = 1/8, reverse = FALSE) +
geom_text(stat = "stratum", label.strata = TRUE, reverse = FALSE, size = 2.5) +
scale_x_continuous(breaks = 1:3, labels = c("phase", "causes", "total crashes")) +
coord_flip() +
labs(title = "Crash Cause", caption = "Source: bit.ly/KIB_PlaneCrashes") +
theme(panel.background = element_blank(),
plot.title = element_text(size = 13,
family = "Georgia",
face = "bold",
lineheight = 1.2,
vjust = -3,
hjust = 0.05),
plot.caption = element_text(size = 5,
hjust = 0.99, family = "Georgia"),
axis.text = element_text(family = "Georgia"),
axis.text.x = element_blank(),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank())
In my opinion, a cleaner way to produce a Sankey diagram is with the package flipPlotpackage.
devtools::install_github(Displayr/flipPlots)
library(flipPlots)# reorder the df
crash$total_crashes <- rep("YES", 25)# plot
SankeyDiagram(crash[, -4], link.color = "Source", label.show.varname = FALSE, weights = crash$freq)
I would probably just rotate the image and labels and change the yes to “Total Crashes (427)”
Gender Gap
This visualization depicts the salary gap between males and females by industry in the UK with the mean salary of each position within a category. We can use group_by() and summarize_at() to create a new variable for each category and then use facet_wrap() . Since positions only belong to one category you need to set scales = "free_x" for missing observations.
gendergap <- read.csv("gendergap.csv")# gather the dataset
tidy_gap <- gendergap %>%
gather(key = sex, value = salary, -title, -category)category_means <- tidy_gap %>%
group_by(category) %>%
summarize_at(vars(salary), mean)tidy_gap %>% ggplot(aes(x = title, y = salary, color = sex)) +
facet_grid(. ~ category, scales = "free_x", space = "free") +
geom_line(color = "white") +
geom_point() +
scale_color_manual(values = c("#F49171", "#81C19C")) +
geom_hline(data = category_means, aes(yintercept = salary), color = "white", alpha = 0.6, size = 1) +
theme(legend.position = "none",
panel.background = element_rect(color = "#242B47", fill = "#242B47"),
plot.background = element_rect(color = "#242B47", fill = "#242B47"),
axis.line = element_line(color = "grey48", size = 0.05, linetype = "dotted"),
axis.text = element_text(family = "Georgia", color = "white"),
axis.text.x = element_text(angle = 90),
# Get rid of the y- and x-axis titles
axis.title.y=element_blank(),
axis.title.x=element_blank(),
panel.grid.major.y = element_line(color = "grey48", size = 0.05),
panel.grid.minor.y = element_blank(),
panel.grid.major.x = element_blank(),
strip.background = element_rect(color = "#242B47", fill = "#242B47"),
strip.text = element_text(color = "white", family = "Georgia"))
One thing that I’m not sure how to handle is the spacing between each of the variables on the x-axis. Since there is a different number of variables for each facet it would be nice if one could specify they want equal spacing along the x-axis as an option in the facet_wrap(); however, I don’t think it’s possible (if you know a workaround please leave a comment!).
## EDIT SEPTEMBER 5 2018
Thanks to help from the ever so helpful community at StackOverflow I learned it was facet_grid() and not facet_wrap() that I needed. Bonus points if anyone figures out how to manually adjust the distance between each of the categories in the x-axis (e.g. minimize the spacing).
That’s all for me, it’s been fun doing this series and I hope you’ve enjoyed!
