Day 05: Ranking

Code 
# Load required packages
library(owidapi)      # for data retrieval from OWID
library(stringr)      # for string manipulation
library(data.table)   # for data manipulation
library(ggplot2)      # for visualization
library(ggtext)       # for text formatting
library(paletteer)    # for color palettes
library(scales)       # for axis formatting
library(dplyr)        # for rowwise not imported by ggsankey
library(gghighlight)  # for highlighting certain countries
library(showtext)     # for fonts
library(ggsankey)     # for Sankey diagram


# For this visualization, we'll focus on maize production only
dat <- owid_get('maize-production')
colnames(dat)[4] <- "tonnes"
setDT(dat)

# Remove non-country entities (regions, income groups, etc.)
entity_to_rm <- c(
    'World',
    'Americas (FAO)',
    'High-income countries',
    'North America',
    'Upper-middle-income countries',
    'Asia',
    'Eastern Asia (FAO)',
    'China (FAO)',
    'Europe (FAO)',
    'South America (FAO)',
    'South America',
    'Lower-middle-income countries',
    'European Union (27) (FAO)',
    'European Union (27)',
    'Africa (FAO)',
    'Africa',
    'Eastern Europe (FAO)',
    'Net Food Importing Developing Countries (FAO)',
    'Low Income Food Deficit Countries (FAO)',
    'Least Developed Countries (FAO)',
    'Southern Europe (FAO)',
    'Net Food Importing Developing Countries (FAO)',
    'Low Income Food Deficit Countries (FAO)',
    'Least Developed Countries (FAO)',
    'Southern Europe (FAO)',
    'South-eastern Asia (FAO)',
    'Central America (FAO)',
    'Southern Asia (FAO)',
    'Land Locked Developing Countries (FAO)',
    'Eastern Africa (FAO)',
    'Western Europe (FAO)',
    'Low-income countries',
    'Southern Africa (FAO)',
    'Western Africa (FAO)',
    'Western Asia (FAO)',
    'Middle Africa (FAO)',
    'Northern Africa (FAO)',
    'Small Island Developing States (FAO)',
    'Caribbean (FAO)',
    'Oceania (FAO)',
    'Central Asia (FAO)',
    'Northern Europe (FAO)',
    'Belgium-Luxembourg (FAO)',
    'Asia (FAO)',
    'Northern America (FAO)',
    'Europe',
    'Micronesia (FAO)'
)
dat <- dat[!entity_name %in% entity_to_rm]

# Keep only data every 10 years for clearer visualization of trends
dat <- dat[year %% 10 == 0]

# Add new en variable since ggsankey transforms data
dat$en <- dat$entity_name

# Identify the top 10 countries by total maize production
dat_top10 <- dat[,
    .(tonnes = sum(tonnes, na.rm = TRUE)),
    by = .(entity_name)
][order(-tonnes)][1:10]
entity_top10 <- dat[,
    .(tonnes = sum(tonnes, na.rm = TRUE)),
    by = .(entity_name)
][order(-tonnes)][1:10][, entity_name]


dat_top10 <- dat[entity_name %in% entity_top10]

# Create highlight grouping variable using data.table
dat_top10[, highlight_group := ifelse(
    entity_name %in% c("Argentina", "Brazil", "China", "India", 'Mexico', 'United States'),
    entity_name,
    "Other"
)]

# Arrange and factor by highlight group
setorder(dat_top10, tonnes)

# Factorize highlight group
dat_top10[, highlight_group := factor(highlight_group, levels = c("Argentina", "Brazil", "China", "India", "United States", "Mexico", "Other"))]

# Get the maximum year for proper positioning
max_year <- max(dat_top10$year)

# Between band spaces
space <- 5e6

# Color palette
pal <- c('#778BD0FF', '#E2D76BFF', '#95BF78FF', '#4E6A3DFF', '#5F4F38FF', '#D88782FF', '#D3D3D3')

# Fonts
font_add_google("Nunito", "nunito")

# Create labels, Calculate reversed positions against total
dat_labels <- dat_top10[year == max(year)][
  order(tonnes)  # Equivalent to arrange(desc(tonnes))
][, `:=`(
  total_production = sum(tonnes),
  cumulative_tonnes = cumsum(tonnes),
  band_index = .I
)][, `:=`(
  # Reverse by subtracting from total
  label_y = cumulative_tonnes - (tonnes / 2) + (band_index - 1) * space,
  label_x = year + 0.5
)][highlight_group != "Other"]  # Filter out "Other" category

# Create the visualization with labels
g <- ggplot() +
    geom_sankey_bump(
        data = dat_top10,
        aes(
            x = year,
            y = tonnes,
            node = entity_name,
            fill = highlight_group,
            value = tonnes,
            label = entity_name
        ),
        space = 5e6,
        alpha = 0.7,
        type = 'alluvial'
    ) +
     # Add country labels at end points
    geom_text(
        data = dat_labels,
        aes(x = label_x, y = label_y, label = highlight_group, color = highlight_group),
        hjust = 0,  # Left-align text
        vjust = 0.5,
        size = 4,
        fontface = "bold",
        show.legend = FALSE
    ) +
    scale_fill_manual(
        values = pal, guide = "none") +
    scale_color_manual(
        values = pal, guide = "none"
    ) +
    scale_y_continuous(
        labels = scales::label_number(scale_cut = scales::cut_short_scale()),
        expand = c(0, 0)
    ) +
    coord_cartesian(clip = "off") +
    labs(
        x = "",
        y = "",
        caption = "Data source: Our World in Data - Agricultural Production | Visualization: @gnoblet"
    ) +
    theme_bw() +
    theme(
        text = element_text(family = "Nunito"),
        legend.position = "right",
        plot.title = element_text(face = "bold", size = 20),
        plot.subtitle = element_text(size = 16),
        plot.caption = element_text(size = 12),
        panel.grid.minor = element_blank(),
        panel.grid.major.x = element_blank(),
        panel.grid.major.y = element_blank(),
        axis.text = element_text(size = 12),
        axis.text.y=element_blank(),
        axis.ticks.y=element_blank(),
        plot.margin = margin(20, 80, 20, 20),
        panel.border = element_blank()
    ) +
    geom_textbox(aes(
            label = "<span style='font-size:20pt; font-family:Nunito; font-weight:bold;'>Global Maize Production Trends (1970-2020)</span>"),
        x = 1969,
        y = 890e6,
        fill = NA,
        hjust = 0,
        fontface = 'bold',
        width = unit(450, "pt"),
        box.colour = NA) +
    geom_textbox(aes(
            label = "<span style='font-size:16pt; font-family:Nunito; line-height:1.8;'> Global maize production for the top 10 producing countries grew from 197M to 897M tonnes in 50 years. Ranking did not change for the top 3 countries as production is largely dominated by the United States, China, and Brazil.</span>"),
        x = 1969,
        y = 850e6,
        fill = NA,
        hjust = 0,
        vjust = 1,
        lineheight = 1.9,
        width = unit(360, "pt"),
        box.colour = NA)


# Save the plot at high resolution
ggsave(
  "day_05.png",
  plot = g,
  height = 9,
  width = 10,
  dpi = 600,
  type = "cairo-png"
)

Final Plot

Notes

This visualization uses a Sankey bump chart to track maize production for the top 10 producing countries from 1970 to 2020. The chart highlights both ranking changes and production volume changes simultaneously.

Data source: Our World in Data - Agricultural Production

Tools used:

  • owidapi (for data retrieval from Our World in Data)
  • data.table (for efficient data manipulation)
  • ggsankey (for creating the Sankey bump chart)
  • ggplot2 (for general plotting framework)
  • paletteer (for color palettes)

The process involved retrieving agricultural production data, filtering for maize production only, removing non-country entities, identifying the top 10 producing countries, and visualizing their production changes over decades. The thickness of each flow represents production volume, making it easy to observe both relative rankings and absolute changes simultaneously.