Loading packages

library(foreign)
library(tidyverse)
library(countrycode)
library(gganimate)
library(scales)
library(RColorBrewer)

Preparing the data

# Reading in the data

data = read.dta("Billionaires1996-2014.dta")

# Adding a variable for countryname using the 3 digit country code

data = data %>%
  mutate(countryname = countrycode(data$countrycode, "iso3c", "country.name")) %>%
  mutate(countryname = case_when(countrycode == "ROM" ~ "Romania",
                                 countrycode == "Taiwan" ~ "Taiwan", 
                                 countrycode == "CHI" ~ "Channel Islands",
                                 countrycode == "HKG" ~ "Hong Kong",
                                 TRUE ~ .$countryname))

# Adding a variable for continent

data = data %>%
  mutate(continent = countrycode(data$countryname, "country.name", "continent")) %>%
  select(-c(citizenship, company))

Figure 1 - Number of billionaires

data %>% 
  group_by(year) %>%
  count() %>%
  ggplot(aes(x = year, y = n)) + 
  geom_line(size=1) +
  expand_limits(y = 0) + 
  labs(x = "",
       y = "Billionaires",
       title = "Number of Billionaires in the World (1996-2016)") +
  geom_point(aes(x= 2001, y=538), colour="#e41a1c", size = 2) +
  geom_point(aes(x= 2008, y=1124), colour="#e41a1c", size = 2) +
  annotate("text", x= 2001, y= 750, label = "2001:\nDot-com bubble burst") +
  annotate("text", x= 2008, y=1330, label = "2008:\nOnset of financial crisis") +
  theme_classic() +
  theme(axis.title.y = element_text(margin = margin(t = 0, r = 15, b = 0, l = 0)),
        plot.title = element_text(size = 16))

Figure 2 - Billionaires by continent

continent_data = data %>%
  filter(!is.na(continent))
continent_data$continent = factor(continent_data$continent, levels=c("Americas","Asia","Europe", "Oceania", "Africa"))
continent_data %>% 
  group_by(year, continent) %>%
  filter(!is.na(continent)) %>%
  count() %>%
  ggplot(aes(x = year, y = n, colour = continent)) + 
  geom_line(size = 1) +
  labs(x = "",
       y = "Billionaires",
       title = "Billionaires by Continent (1996-2016)",
       colour = "") +
  theme_classic() +
  theme(axis.title.y = element_text(margin = margin(t = 0, r = 15, b = 0, l = 0)),
        plot.title = element_text(size = 16)) +
  scale_colour_brewer(type = "qual", palette = 6)

# I wanted to label each line directly instead of using a legend; however, since
# the Oceania and Africa lines are so close together, the labels overlapped.

Figure 3 - Billionaires in emerging and advanced economies

data %>%
  mutate(north = recode_factor(north,
                              "0" = "Emerging economies",
                              "1" = "Advanced economies")) %>%
  group_by(year, north) %>%
  summarise(totalnetworth = sum(realnetworth, na.rm = TRUE)) %>%
  ggplot(aes(x = year, y = totalnetworth, colour = north)) +
  geom_line(size = 1) +
  expand_limits(y = 0) +
  labs(x = "",
       y = "",
       title = "Total Net Worth of Billionaires in Emerging & Advanced Economies",
       subtitle = "Measured in billions, 1996 US dollars",
       colour = "") +
  theme_classic() +
  scale_colour_brewer(type = "qual", palette = 6,
                      guide = guide_legend(reverse=TRUE)) +
  scale_y_continuous(labels = dollar_format()) + 
  theme(plot.title = element_text(size = 16))

# I also wanted to label each line directly instead of using a legend; however,
# I chose not to do this because I wanted to ensure consistency with the other
# graphs + fitting in the full labels ("advanced economies" and "emerging
# economies") distorted the x-axis significantly.

Figure 4 - Bilionaires by industry

x = data %>%
  filter(year %in% c(2000, 2004, 2008, 2012, 2016)) %>%
  group_by(year, IndustryAggregates) %>%
  filter(!is.na(IndustryAggregates)) %>%
  summarise(count = n()) %>%
  spread(IndustryAggregates, count) %>%
  mutate(total = sum(other, `Resource Related`, `New Sectors`, `Non-Traded Sectors`, `Financial`, `Traded Sectors`)) %>%
  summarise(Other = other/total, 
            `Resource Related` = `Resource Related`/total, 
            `New` = `New Sectors`/total, 
            `Non-Traded` = `Non-Traded Sectors`/total, 
            `Financial` = `Financial`/total, 
            `Traded` = `Traded Sectors`/total) %>%
  gather("industry", "percentage", 2:7)

x$industry <- factor(x$industry, levels = c("Other", "New", "Resource Related", "Traded", "Non-Traded", "Financial"))

x %>% 
  ggplot() +
  geom_bar(aes(x = year, y = percentage, fill = industry), stat = "identity") +
  scale_x_continuous(breaks=seq(2000,2016,4)) +
  scale_y_continuous(labels = percent_format()) +

# Position_stack() puts the text annotations in the middle of each bar segment
  
  geom_text(aes(x = year, y = percentage, label = paste0(round(percentage * 100), "%"), group = industry),
                  position = position_stack(vjust = .5), size = 3.5) +
  theme_classic() +
  labs(x = "",
       y = "Percentage of Billionaires",
       title = "Billionaires by Industry",
       fill = "Industry") +
  theme(plot.title = element_text(size = 16)) +
  scale_fill_brewer(type = "qual", palette = 4, direction = -1)

Figure 5 - Billionaires by country

# Creating a rank variable for the top 10 countries each year

x = data %>%
  group_by(year) %>%
  count(countryname) %>%
  mutate(rank = rank(-n, ties.method = "first")) %>%
  group_by(countryname) %>%
  filter(rank <= 10)

# The ColorBrewer palette I'm using doesn't have enough unique colours (I need
# 19 since that's how many countries make the top 10 at least once). This code
# creates enough new shades from the existing palette.

colourCount = length(unique(x$countryname))
getPalette = colorRampPalette(brewer.pal(colourCount,"Set1"))

anim = x %>% 
  ggplot(aes(rank, group = countryname, fill = as.factor(countryname))) +

# I'm using geom_title because it slides more smoothly than a regular bar graph.
  
  geom_tile(aes(y = n/2, 
                height = n, 
                width = 0.9), alpha = 0.8) +
  geom_text(aes(y = 0, label = paste(countryname, " ")), vjust = 0.2, hjust = 1) +
  geom_text(aes(y = n, label = paste(n, " ")), hjust = -0.2) +
  coord_flip(clip = "off", expand = FALSE) +
  scale_x_reverse() +

# Closest_state shows the year of the current frame
  
  labs(title='Countries with the most billionaires in {closest_state}', x = "", y = "") + 
  theme(legend.position = "none",
        axis.ticks.y = element_blank(),
        axis.text.y  = element_blank(),
        axis.ticks.x = element_blank(),
        axis.text.x  = element_blank(),
        panel.grid.major=element_blank(),
        panel.grid.minor=element_blank(),
        plot.margin = margin(2,2, 2, 4, "cm"),
        panel.background = element_blank(),
        plot.title = element_text(size = 16)) +
  scale_fill_manual(values = getPalette(colourCount)) +
  scale_colour_manual(values = getPalette(colourCount)) +
  transition_states(year, transition_length = 10, state_length = 10) +
  ease_aes('cubic-in-out') 

animate(anim, nframes = 80, fps = 4, end_pause = 20)