Example analysis and visualizations of World Bank data for ASEAN GDP indicators.

asean-gdp-comparisons.R

if (!require("pacman")) install.packages("pacman")
pacman::p_load("dplyr", "ggplot2", "readr", "reticulate", "showtext")

# Utility function to save ggplots
save_image <- function(filename, plot, width = 7, height = 7) {
    ggsave(
        sprintf("%s.png", filename),
        plot,
        width = width, height = height,
        units = "in", dpi = 96
    )
}
# ggplot theme and text settings
plot_settings <- list(
    labs(
        caption = "Data: World Bank | github.com/emordonez"
    ),
    theme_minimal(),
    theme(
        legend.position = "none",
        plot.margin = margin(10, 10, 10, 10, "pt"),
        text = element_text(family = "Open Sans"),
        plot.title = element_text(face = "bold"),
        plot.subtitle = element_text(face = "plain"),
        plot.caption = element_text(face = "italic"),
        axis.title.x = element_text(margin = margin(7, 0, 0, 0, "pt")),
        axis.title.y = element_text(margin = margin(0, 7, 0, 0, "pt"))
    )
)
font_add_google("Open Sans", "Open Sans")
showtext_auto(enable = TRUE)

# Read data
source_python("data.py")
df <- read_data("./world-bank-data.csv")
df <- df %>%
    mutate(
        country = case_when(
            country == "Brunei Darussalam" ~ "Brunei",
            country == "Hong Kong SAR, China" ~ "Hong Kong",
            country == "Lao PDR" ~ "Laos",
            country == "Korea, Rep." ~ "South Korea",
            TRUE ~ country
        ),
        gdp_constant_usd = gdp_constant_usd / 1e3,
        gdp_per_capita = gdp_per_capita / 1e3
    )
asean_names <- c(
    "Brunei", "Cambodia", "Indonesia", "Laos", "Malaysia",
    "Myanmar", "Philippines", "Singapore", "Thailand", "Vietnam"
)
other_country_names <- c(
    "Australia", "China", "European Union", "Hong Kong",
    "India", "Japan", "South Korea", "United States"
)
asean_countries <- df %>% filter(country %in% asean_names)
other_countries <- df %>% filter(country %in% other_country_names)
asean <- df %>% filter(country == "ASEAN")
world <- df %>% filter(country == "World")


#'
#' PLOT 1: Real GDP per capita among ASEAN countries
#' 
world_gdp_pc <- tail(world$gdp_per_capita, n = 1)
viz_rgdp_pc_2019 <- asean_countries %>%
    filter(year == 2019) %>%
    select(country, gdp_per_capita) %>%

    ggplot(aes(x = reorder(country, gdp_per_capita), y = gdp_per_capita)) +
    geom_col(aes(alpha = 0.75, fill = country)) +
    geom_text(
        aes(
            label = sprintf("%0.1f", gdp_per_capita),
            y = gdp_per_capita,
            hjust = -0.25
        ),
        size = 4.5
    ) +    
    geom_hline(yintercept = world_gdp_pc) +
    annotate(
        "text",
        x = "Vietnam",
        y = 20,
        label = sprintf("Global average, %0.1f", world_gdp_pc),
        size = 4
    ) +
    labs(
        title = "Real GDP per capita among ASEAN countries",
        subtitle = paste(
            "ASEAN members span lower middle income (Cambodia), upper middle",
            "income \n(Indonesia), and high income economies (Brunei)."
        ),
        x = NULL,
        y = "Real GDP per capita, thousand USD (2010)"
    ) +
    scale_y_continuous(expand = expansion(mult = c(0, 0.08))) +
    plot_settings +
    coord_flip()
save_image("1-rgdp-per-capita-2019", viz_rgdp_pc_2019)


#' 
#' PLOT 2: Average annual growth in real GDP since 2000
#' 
avg_annual_growth_rates <- asean %>%
    rbind(other_countries) %>%
    filter(year == 2000 | year == 2019) %>%
    select(country, year, gdp_constant_usd) %>%
    group_by(country) %>%
    mutate(
        avg_annual_growth = (gdp_constant_usd[year == 2019] / gdp_constant_usd[year == 2000])^(1 / 19) - 1
    ) %>%
    filter(year == 2019) %>%
    select(country, avg_annual_growth) %>%
    ungroup()

viz_avg_annual_growth_rates <- avg_annual_growth_rates %>%
    mutate(country = case_when(
        country == "European Union" ~ "EU",
        country == "United States" ~ "US",
        TRUE ~ country
    )) %>%

    ggplot(aes(x = reorder(country, -avg_annual_growth), y = avg_annual_growth)) +
    geom_bar(aes(fill = country, alpha = 0.95), stat = "identity", width = 1, color = "white") +
    geom_label(
        aes(
            label = sprintf("%s\n%0.1f%%", country, avg_annual_growth * 100)
        ),
        fill = alpha(c("white"), 1),
        y = 0.1035,
        size = 4.5,
        label.size = NA
    ) +
    labs(
        title = "Average annual growth in real GDP since 2000",
        subtitle = paste(
            "Collectively, ASEAN's economic growth is outpaced in Asia only by",
            "China and India."
        ),
        x = NULL,
        y = NULL
    ) +
    plot_settings +
    theme(
        panel.grid.major = element_line(color = "grey85"),
        axis.text = element_blank()
    ) +
    coord_polar()
save_image("2-rgdp-growth-since-2000", viz_avg_annual_growth_rates)


#'
#' PLOT 3: Asia's growth in income since 1967
#' 
gdp_since_1967 <- asean %>%
    rbind(other_countries) %>%
    filter(country %in% c("ASEAN", "China", "India") & year >= 1967) %>%
    select(country, year, gdp_per_capita)

china_gdp_pc_2002 <- tail(
    filter(gdp_since_1967, country == "China" & year == 2002)$gdp_per_capita,
    n = 1
)
viz_gdp_since_1967 <- gdp_since_1967 %>%
    ggplot(aes(x = year, y = gdp_per_capita)) +
    geom_line(aes(color = country), size = 1) +
    annotate(
        "rect",
        xmin = 1996, xmax = 1999,
        ymin = 0, ymax = Inf,
        fill = "grey", alpha = 0.35
    ) +
    annotate(
        "point",
        x = 2002,
        y = china_gdp_pc_2002,
        color = "black"
    ) +
    annotate(
        "text",
        x = 1987,
        y = 3.5,
        label = "1997 Asian financial crisis",
        size = 4
    ) +
    annotate(
        "text",
        x = 2010,
        y = 2,
        label = "China joins the WTO,\nDec 2001",
        size = 4
    ) +
    labs(
        title = "Asia's growth in income since 1967",
        subtitle = paste(
            "Measured by GDP per capita, China has overtaken ASEAN in prosperity",
            "only within \nthe past 15 years. ASEAN was most affected by the",
            "Asian financial crisis but has \nrecovered to pre-crisis levels",
            "of growth."
        ),
        x = "Year",
        y = "Real GDP per capita, thousand USD (2010)"
    ) +
    plot_settings +
    theme(
        legend.title = element_blank(),
        legend.position = c(0, 1),
        legend.justification = c(0, 1),
        legend.text = element_text(size = 10)
    )
save_image("3-asian-income-since-1967", viz_gdp_since_1967)


#'
#' PLOT 4: GDP growth among ASEAN countries since 2000
#' 
asean_since_2000 <- asean_countries %>%
    filter(year >= 2000) %>%
    mutate(
        income = case_when(
            code %in% c("SGP", "BRN") ~ "High income (Singapore, Brunei)",
            code %in% c("MYS", "THA", "IDN") ~ "Upper middle income (Malaysia, Thailand, Indonesia)",
            TRUE ~ "Lower middle income (Philippines, Vietnam, Laos, Myanmar, Cambodia)"
        )
    ) %>%
    select(country, year, gdp_growth, income)

asean_avg_rate <- tail(
    filter(avg_annual_growth_rates, country == "ASEAN")$avg_annual_growth,
    n = 1
) * 100
viz_asean_since_2000 <- asean_since_2000 %>%
    ggplot(aes(x = year, y = gdp_growth * 100, color = country)) +
    geom_line(size = 0.5) +
    geom_point(size = 0.75) +
    geom_hline(yintercept = asean_avg_rate, linetype = "dashed") +
    annotate(
        "rect",
        xmin = 2000, xmax = 2001,
        ymin = -Inf, ymax = Inf,
        fill = "grey", alpha = 0.35
    ) +
    annotate(
        "rect",
        xmin = 2007, xmax = 2009,
        ymin = -Inf, ymax = Inf,
        fill = "grey", alpha = 0.35
    ) +
    annotate(
        "text",
        x = 2017,
        y = 10,
        label = sprintf("ASEAN average, %0.1f%%", asean_avg_rate),
        size = 3.5
    ) +
    labs(
        title = "GDP growth among ASEAN countries since 2000",
        subtitle = paste(
            "The aftermath of the 1997 financial crisis and the 2008 crisis",
            "impacted the more \ndeveloped ASEAN countries harder. The less",
            "developed countries consistently \nhave above average growth."
        ),
        x = "Year",
        y = "Annual percent growth of real GDP (2010 USD)"
    ) +
    plot_settings +
    theme(
        strip.text = element_text(size = 10, face = "bold"),
        legend.title = element_blank(),
        legend.position = "bottom"
    ) +
    facet_wrap(
        ~factor(income, levels = c(
            "High income (Singapore, Brunei)",
            "Upper middle income (Malaysia, Thailand, Indonesia)",
            "Lower middle income (Philippines, Vietnam, Laos, Myanmar, Cambodia)")
        ),
        ncol = 1
    )
save_image("4-asean-growth-since-2000", viz_asean_since_2000, height = 11.5)

data.py

import pandas as pd


def read_data(file):
    data = pd.read_csv(file)
    countries = [country for country in set(data['Country Name'])]
    asean_codes = [
        'BRN', 'KHM', 'IDN', 'LAO', 'MYS',
        'MMR', 'PHL', 'SGP', 'THA', 'VNM'
    ]
    # Indicator variables, see CSV metadata
    gdp_indicator = 'GDP (constant 2010 US$)'
    population_indicator = 'Population, total'

    df_list = []
    for country in countries:
        df = data[data['Country Name'] == country]

        gdp = df[df['Indicator Name'] == gdp_indicator]
        gdp = pd.melt(
            gdp,
            id_vars=['Country Name', 'Country Code'],
            # Looking at 1960-2019, 2020 data not in yet
            value_vars=gdp.columns[3:-1],
            var_name='year',
            value_name='gdp_constant_usd'
        )

        population = df[df['Indicator Name'] == population_indicator]
        population = pd.melt(
            population,
            id_vars=['Country Name', 'Country Code'],
            value_vars=population.columns[3:-1],
            var_name='year',
            value_name='population'
        )

        df = gdp.merge(population, on=['Country Name', 'Country Code', 'year'])
        df.rename(
            columns={'Country Name': 'country', 'Country Code': 'code'},
            inplace=True
        )
        df_list.append(df)


    def format_cols(df):
        df['year'] = pd.to_numeric(df.year, downcast='integer')
        df['gdp_constant_usd'] = pd.to_numeric(
            df.gdp_constant_usd, errors='coerce'
        )
        df['population'] = pd.to_numeric(
            df.population, errors='coerce', downcast='integer'
        )
        # NaN can be safely filled with 0
        # Rows with NaN GDP or population will be dropped anyway
        df['gdp_growth'] = df.gdp_constant_usd.pct_change().fillna(0)
        df['pop_growth'] = df.population.pct_change().fillna(0)

        df.dropna(inplace=True)
        df['gdp_per_capita'] = df.gdp_constant_usd / df.population
        df['gdp_pc_growth'] = df.gdp_per_capita.pct_change().fillna(0)

        return df


    for df in df_list:
        df = format_cols(df)
    df = pd.concat(df_list, ignore_index=True)

    # Aggregate ASEAN members onward from the year they joined ASEAN
    asean = df[df.code.isin(asean_codes)]
    # 1967: Indonesia, Malaysia, Philippines, Singapore, Thailand
    asean_67 = asean.loc[asean.code.isin(['IDN', 'MYS', 'PHL', 'SGP', 'THA'])] \
                    .loc[asean.year >= 1967]
    # 1984: Brunei
    asean_84 = asean.loc[asean.code == 'BRN'] \
                    .loc[asean.year >= 1984]
    # 1995: Vietnam
    asean_95 = asean.loc[asean.code == 'VNM'] \
                    .loc[asean.year >= 1995]
    # 1997: Laos, Myanmar
    asean_97 = asean.loc[asean.code.isin(['LAO', 'MMR'])] \
                    .loc[asean.year >= 1997]
    # 1999: Cambodia
    asean_99 = asean.loc[asean.code == 'KHM'] \
                    .loc[asean.year >= 1999]

    asean = pd.concat([
        asean_67,
        asean_84,
        asean_95,
        asean_97,
        asean_99
    ], ignore_index=True)
    asean['country'] = 'ASEAN'
    asean['code'] = 'SEA'
    asean = asean.groupby(['country', 'code', 'year'], as_index=False) \
                .agg({'gdp_constant_usd': 'sum', 'population': 'sum'}) \
                .pipe(format_cols)

    return pd.concat([df, asean], ignore_index=True)

world-bank-data.csv