Creating The Economist plot style with the Governism Index from JOTA

governismo_partidos.csv

satuRday-01.R

# Creating The Economist plot style with the Governism Index from JOTA

# 1 What am I gonna do?
## I will try to replicate The Economist plot sytle using only R packages. The data I am using is a measure calculated every week by the JOTA Labs team at https://jota.info, which tells us about the governism attachment among members of the Brazilian Congress.

# 2 Why we need to use R anyway?
## R is just a data visualization tool, but it is free and programming oriented. Therefore, you just have to build the plot style once, and you can recreate it as many as you want. In addition, R has lots of great packages, one of them is named ggplot2, which has many varieties of supporting elements that you can use to create charming visualization.

# 3 Preparation
## Let's load the required libraries

library(tidyverse)
library(scales)
library(ggrepel)
library(ggthemes)
library(grid)
options(scipen = 100) # to set the scientific notation


# 4 Data preparation

## The data with the governism index and others importante parameters, as the size of the caucus and the number of votes cast in the House, is read right from the gist repository.

library(readr)

data <-
  read_csv(
    "https://gist.githubusercontent.com/danielmarcelino/bebe807acaeb8facf2ae2bfcf284e381/raw/dcd13ad53fcff2893fcdad5db47922ee85c89e2f/governismo_partidos.csv"
  )



glimpse(data)


# 5 Making the Scatter Plot
## First, I plot the data as scatter plot, with the size of the dot indicating the caucus size.

g <- data %>%
  ggplot() +
  geom_point(
    aes(votes, governismo, size = caucus),
    color = "#2fc1d3",
    alpha = .5,
    show.legend = F
  )
g




## Next, I will smooth the data using the log10 scale

g <- g +
  scale_x_continuous(
    trans = "log10",
    expand = c(0.05, 0),
    labels = number_format(big.mark = ","),
    limits = c(100, 20000)
  )
g



## Now, I rescale the size of the dot, so it will look nicely

g <- g +
  scale_size_continuous(range = c(2, 8), guide = F)
g



## I also have to scale the y axis

g <- g +
  scale_y_continuous(
    breaks = seq(0, 1, .2),
    limits = c(0, 1),
    position = "right",
    expand = c(0, 0),
    labels = scales::percent
  )
g



# 6 Plot Theme
## It's time to clean the background and the grid with this small hack

g <- g +
  theme_economist() +
  theme(
    panel.grid.major.y = element_line(colour = "gray80"),
    panel.grid.minor.y = element_blank(),
    panel.grid.major.x = element_blank(),
    panel.grid.minor.x = element_blank(),
    panel.background = element_blank(),
    plot.background = element_blank()
  )

g



## Now adjust the tick mark and the position of the y-axis text

g <- g +
  theme(axis.ticks.length.x = unit(2, "mm"),
        axis.text.y = element_text(vjust = 0))
g




# 7 Add title and annotations
## I will remove the y-axis title and rename the x-axis title. Also I add the title and caption information

g <- g +
  theme(
    axis.title.y = element_blank(),
    axis.title.x = element_text(colour = "black", size = 12),
    plot.title = element_text(size = 18, face = "bold", color = "black"),
    plot.subtitle = element_text(size = 14, color = "black"),
    plot.caption = element_text(
      size = 10,
      color = "gray30",
      hjust = 0
    )
  ) +
  labs(
    title = "Like a concert without a crowd",
    subtitle = "Governism and roll call votes, August/2020",
    caption = "Source: http://jota.info",
    x = "Number of votes cast in roll calls, log scale"
  )

g




# Next I can add the annotation of governism level and caucus size

g <- g +
  geom_hline(
    aes(yintercept = .74),
    color = "#edb0ad",
    linetype = "dashed",
    size = .85
  ) +
  geom_text(
    aes(100, .765, label = "Average governism level"),
    color = "#e07b78",
    hjust = "left",
    size = 4
  ) +
  geom_text(
    aes(100, .225, label = "Circle size = Caucus size, 2020"),
    color = "#8e9093",
    hjust = "left",
    size = 4
  )
g




# 8 Highligt selected parties
## I'll highlight and add annotation for selected parties. Since the annotation line is in elbow shape, we need to be creative. Perhaps someone knows a package that can directly make an elbow annotation

### First, I add the smallest political party REDE

data_rede <- data %>% filter(legislator_party == "REDE")

g <- g +
  geom_text_repel(
    aes(votes, governismo, label = legislator_party),
    data = data_rede,
    nudge_x = -.15,
    direction = "x"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_rede,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )

g



### Next, I add some of the major parties PSL and DEM

data_dem <-
  data %>% filter(legislator_party %in% c("DEM", "PSL"))


g <- g +
  geom_segment(aes(
    x = votes,
    xend = votes,
    y = governismo,
    yend = governismo - .21
  ),
  data = data_dem) +
  geom_text_repel(
    aes(votes, governismo - .21, label = legislator_party),
    data = data_dem,
    nudge_x = -.15,
    direction = "x"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_dem,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g



### Next, we add PSDB

data_psdb <- data %>% filter(legislator_party == "PSDB")

g <- g +
  geom_text_repel(
    aes(votes, governismo, label = legislator_party),
    data = data_psdb,
    nudge_y = -.25,
    direction = "y"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_psdb,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g


### Next, we add some of the opposition parties

data_pt <-
  data %>% filter(legislator_party == "PT")


g <- g +
  geom_segment(aes(
    x = votes,
    xend = votes,
    y = governismo,
    yend = governismo + .1
  ),
  data = data_pt) +
  geom_text_repel(
    aes(votes, governismo + .1, label = legislator_party),
    data = data_pt,
    nudge_x = -.25,
    direction = "x"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_pt,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g




data_psol <- data %>% filter(legislator_party == "PSOL")

g <- g +
  geom_text_repel(
    aes(votes, governismo, label = legislator_party),
    data = data_psol,
    nudge_x = .25,
    direction = "x"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_psol,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g




### We can add a small but very loyal party

data_patri <- data %>% filter(legislator_party == "PATRI")

g <- g +
  geom_text_repel(
    aes(votes, governismo, label = legislator_party),
    data = data_patri,
    nudge_y = -.12,
    direction = "y"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_patri,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g




## Finally, we add PP, the party of the government new floor leader

data_pp <-
  data %>% filter(legislator_party == "PP")

g <- g +
  geom_segment(aes(
    x = votes,
    xend = votes,
    y = governismo,
    yend = governismo - .075
  ),
  data = data_pp) +
  
  geom_text_repel(
    aes(votes, governismo - .075, label = legislator_party),
    data = data_pp,
    nudge_x = -.3,
    direction = "x",
    fontface = "bold"
  ) +
  geom_point(
    aes(votes, governismo, size = caucus),
    data = data_pp,
    shape = 21,
    fill = "#2fc1d3",
    color = "black"
  )
g


# 9 Final Result: the cherry on the top

##  To put the unique red header and red rectangle on the top left side as of The Economist plot, I will use the `grid.rect` twice.


png(
  "the_economist_style_plot.png",
  width = 9,
  height = 6,
  units = "in",
  res = 300
)

print(g)

grid.rect(
  x = 1,
  y = .995,
  hjust = 1,
  vjust = 0,
  gp = gpar(fill = '#e5001c', lwd = 0)
)


grid.rect(
  x = 0.05,
  y = .98,
  hjust = 1,
  vjust = 0,
  gp = gpar(fill = '#e5001c', lwd = 0)
)

dev.off()