Background Traditional demographic indicators offer a limited view of a country’s population structure and may not capture demographic influences on mortality patterns comprehensively. We aimed to bridge this gap by developing novel scalar metrics that condense the information in population pyramids and assess their association with disease specific mortality.
Methods Country specific population pyramids were constructed using the United Nations World Population Prospects 2024, while mortality data for 371 diseases across 180 countries were extracted from the Global Burden of Disease Study 2021. We then developed two metrics: PoPDivergence, which quantifies the difference between a country’s population pyramid and an optimized reference pyramid using Kullback Leibler divergence, and PoPStat, which is the correlation between PoPDivergence and cause specific mortality rates.
Findings Non communicable diseases (NCDs) showed a strong PoPStat of –0.84 (optimized reference: Japan, p<0.001) with mortality concentrated to constrictive pyramids. Communicable, maternal, neonatal, and nutritional diseases had a moderate PoPStat of 0.50 (Singapore, p<0.001) with mortality linked to expansive pyramids. Injuries exhibited a weak PoPStat (0.291, Singapore, < 0.001). In more granular analyses, NCDs like neurological disorders and neoplasms, communicable diseases (CDs) like neglected tropical diseases, and other infections showed a strong PoPStat. Nevertheless, NCDs like diabetes, cirrhosis, and other chronic liver diseases, CDs like respiratory infections and tuberculosis carried a weak PoPStat indicating minimal influence of population pyramid on their mortality. Interpretations This study, its devised metrics, demonstrates the degree to which the mortality of different diseases is bound to the underlying population structure and reveals what type of population pyramids will carry the highest mortality attributed to those diseases.
Funding This study received no external funding.
Keywords Population Pyramids, Demographic Transition, Epidemiology, PopDivergence, PoPStat
Clone the project repository to your local machine:
git clone https://github.com/Buddhi19/Pop_Pyramid.git
cd Pop_PyramidRun the setup script to install dependencies and prepare the workspace:
bash ./setup_environment.shThe following datasets are required for analysis. Download them and place them in their respective folders as outlined below:
- 👶 Age Dataset
- 🦠 Owid COVID Dataset
- 📊 Median Age
- 📈 SDI
- 🌍 HDI
- 💰 GDP per Capita
- 🌡️ Life Expectancy
- 🏙️ Population Density
- ⚰️ Mortality Data
Alternatively, download the pre-structured data folder from Google Drive.
Organize your downloaded data to match the following structure:
📊 Pop_Pyramid
│
├── 📈 ANALYSIS
├── 🔬 ANALYSIS_FOR_OTHER_DISEASES
├── 📁 DATA
│ ├── 🌍 countries
│ ├── 🦠 covid_data_by_country
│ ├── ⚰️ deaths_by_cause
│ ├── ⚰️ deaths_by_cause_per_country
│ ├── 🗂️ owid_covid_data
│ │ └── 📄 owid-covid-data.csv
│ ├── 🗂️ owid_data
│ │ ├── 📄 median-age.csv
│ │ ├── 📄 population-density.csv
│ │ ├── 📄 life-expectancy.csv
│ │ ├── 📄 gdp-per-capita.csv
│ │ ├── 📄 human-development-index.csv
│ │ └── 📄 sdi_data.csv
│ ├── 🗂️ death_data
│ │ └── 📄 deaths_by_cause.csv
│ ├── 👥 population_data_by_country
│ └── 👶👴 population_data_with_age
│ └── 📄 age_data.csv
Set your parameters and run the analysis:
- Population Year: Enter the year for population data (e.g.,
2020). - COVID Data Date: Enter the desired date for COVID data (format:
YYYY-MM-DD, e.g.,2022-04-08). - Plot Population Pyramids: Choose whether to generate visualizations (
y/n).
Run the analysis:
python -m ANALYSIS --py <population year> --cd "<covid data date>" --plot "<y/n>"python -m ANALYSIS --py 2020 --cd "2022-04-08" --plot "y"To analyze death data for all diseases:
python ANALYSIS_FOR_OTHER_DATA/Analysis_of_Death_Data.pyAfter running the analyses, output files and visualizations will be saved in the RESULTS folder. Review these for insights and further interpretation.
If you encounter any issues or have questions, feel free to open an issue on the GitHub repository.