TidyTuesday Week 45: WHO TB Burden Data: Incidence, Mortality, and Population

This week, we explore global tuberculosis (TB) burden estimates from the World Health Organization, using data curated via the getTBinR R package by Sam Abbott. The dataset includes country-level indicators such as TB incidence, mortality, case detection rates, and population estimates across multiple years. These metrics help researchers, public health professionals, and learners understand the scale and distribution of TB worldwide.

Area plot of TB numbers over the years

1. Python code

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker

# Load data
who_tb_data = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-11-11/who_tb_data.csv')

# Select and summarize
df = who_tb_data[["country", "year", "g_whoregion", "c_cdr", "e_mort_num", "e_inc_num", "e_mort_exc_tbhiv_num", "e_mort_tbhiv_num"]]

mort_summary = (
    df.groupby(["g_whoregion", "year"], as_index=False)
      .agg({"e_mort_exc_tbhiv_num": "sum", "e_mort_tbhiv_num": "sum"})
      .rename(columns={
          "e_mort_exc_tbhiv_num": "mort_nonhiv",
          "e_mort_tbhiv_num": "mort_hiv"
      })
)

# Reshape long
mort_summary = mort_summary.melt(
    id_vars=["g_whoregion", "year"],
    value_vars=["mort_nonhiv", "mort_hiv"],
    var_name="mort_type",
    value_name="deaths"
)

# Rename and order mortality types
mort_summary['mort_type'] = mort_summary['mort_type'].map({
    'mort_hiv': 'HIV related Tuberculosis Death',
    'mort_nonhiv': 'Non-HIV related Tuberculosis Death'
})

mort_summary['mort_type'] = pd.Categorical(
    mort_summary['mort_type'],
    categories=[
        'Non-HIV related Tuberculosis Death',
        'HIV related Tuberculosis Death'
    ],
    ordered=True
)

# Determine plot structure
mort_types = mort_summary['mort_type'].unique()
fig, axes = plt.subplots(1, len(mort_types), figsize=(14, 6), sharey=True)

if len(mort_types) == 1:
    axes = [axes]

for ax, mtype in zip(axes, mort_types):
    subset = mort_summary[mort_summary['mort_type'] == mtype]
    pivot_df = subset.pivot_table(
        index='year', columns='g_whoregion', values='deaths', aggfunc='sum'
    ).fillna(0)

    ax.stackplot(
        pivot_df.index,
        pivot_df.T.values,
        labels=pivot_df.columns
    )

    ax.set_title(mtype)
    ax.set_xlabel("Year")
    ax.set_ylabel("Deaths")
    ax.grid(False)
    ax.spines['top'].set_visible(False)
    ax.spines['right'].set_visible(False)
    ax.spines['left'].set_color("darkgrey")
    ax.spines['bottom'].set_color("darkgrey")
    ax.yaxis.set_major_formatter(mticker.StrMethodFormatter('{x:,.0f}'))

fig.legend(
    pivot_df.columns,
    bbox_to_anchor=(0.5, -0.08),
    title="WHO-Defined Region",
    loc="lower center",
    ncol=3,
    frameon=False
)
fig.suptitle("Non-HIV and HIV-related Tuberculosis deaths per WHO-Region", fontsize=14)
plt.tight_layout(rect=[0, 0, 1, 0.95])