🌍 Epidatasets

Sponsored by

AI-powered epidemiological intelligence

A Python library providing unified access to 41 epidemiological data sources from around the world, with a plugin registry, CLI, and optional extras for specialized data.

📋 Table of Contents

• Overview
• Installation
• Quick Start
• Repository Structure
• Available Datasets
  • Global
  • North America
  • South America
  • Europe
  • Africa
  • Asia
  • Oceania
• CLI Usage
• Usage Examples
• Available Sources
• FAQ
• Contributing
• Related Projects
• Citation
• License

🎯 Overview

epidatasets provides:

• Unified interface — A single get_source() API to access 41 data sources worldwide
• Plugin registry — Sources are discovered at runtime via entry_points, making it easy to extend
• Optional extras — Install only the dependencies you need (pip install epidatasets[who,brazil])
• CLI — Command-line tool for listing sources, inspecting metadata, and querying countries
• Caching & rate limiting — Built-in utilities for responsible API usage
• Reproducible research — Standardized access to heterogeneous epidemiological datasets

📦 Installation

From PyPI

pip install epidatasets

With optional extras

# WHO Global Health Observatory data
pip install epidatasets[who]

# Brazilian DATASUS/SINAN data via PySUS
pip install epidatasets[brazil]

# Eurostat EU health statistics
pip install epidatasets[eurostat]

# Climate/environmental data (Copernicus CDS)
pip install epidatasets[climate]

# Google Earth Engine satellite imagery
pip install epidatasets[earthengine]

# Geospatial visualization
pip install epidatasets[geo]

# Plotting & visualization
pip install epidatasets[viz]

# Genomic data (Pathoplexus)
pip install epidatasets[genomics]

# CLI support
pip install epidatasets[cli]

# World Bank indicators
pip install epidatasets[worldbank]

# Install everything
pip install epidatasets[all]

Development installation

git clone https://github.com/fccoelho/epidemiological-datasets.git
cd epidemiological-datasets
pip install -e ".[dev,docs]"

🚀 Quick Start

from epidatasets import get_source, list_sources

# Discover available sources
sources = list_sources()
for name, meta in sorted(sources.items()):
    print(f"{name}: {meta['description']}")

# Get a specific source
paho = get_source("paho")
countries = paho.list_countries()
print(f"PAHO covers {len(countries)} countries")

# Get WHO data (requires: pip install epidatasets[who])
who = get_source("who")
malaria = who.get_indicator(
    indicator="MALARIA_EST_INCIDENCE",
    years=[2020, 2021, 2022],
    countries=["BRA", "IND", "NGA"]
)

# Get OWID COVID-19 data
owid = get_source("owid")
covid = owid.get_covid_data(
    countries=["BRA", "USA", "IND"],
    metrics=["cases", "deaths"]
)

📁 Repository Structure

epidemiological-datasets/
├── src/epidatasets/           # Main Python package
│   ├── __init__.py            # Public API (get_source, list_sources)
│   ├── _base.py               # BaseAccessor ABC
│   ├── _registry.py           # Plugin registry (entry_points)
│   ├── cli.py                 # CLI (typer)
│   ├── newsfeeds/             # RSS/news-feed aggregator + spatiotemporal viz
│   ├── sources/               # 32 data source accessors
│   │   ├── __init__.py
│   │   ├── africa_cdc.py
│   │   ├── cdc_opendata.py
│   │   ├── china_cdc.py
│   │   ├── hk_chp.py
│   │   ├── colombia_ins.py
│   │   ├── copernicus_cds.py
│   │   ├── datasus_pysus.py
│   │   ├── ecdc_opendata.py
│   │   ├── epipulse.py
│   │   ├── eurostat.py
│   │   ├── global_health.py
│   │   ├── google_earth_engine.py
│   │   ├── healthdata_gov.py
│   │   ├── india_idsp.py
│   │   ├── infodengue_api.py
│   │   ├── malaria_atlas.py
│   │   ├── nhs_prescribing.py
│   │   ├── nz_health.py
│   │   ├── oman_moh.py
│   │   ├── opendengue.py
│   │   ├── owid.py
│   │   ├── paho.py
│   │   ├── pakistan_nih.py
│   │   ├── pathoplexus.py
│   │   ├── respicast.py
│   │   ├── rki_germany.py
│   │   ├── srag_opendata_sus.py
│   │   ├── swiss_bag.py
│   │   ├── ukhsa.py
│   │   └── who_ghoclient.py
│   └── utils/                 # Utilities
│       ├── cache.py           # Caching layer
│       ├── rate_limit.py      # API rate limiting
│       ├── geo.py             # Geospatial helpers
│       ├── validation.py      # Data validation
│       └── io.py              # I/O utilities
├── tests/                     # Test suite
│   ├── sources/
│   ├── utils/
│   ├── conftest.py
│   └── ...
├── docs/                      # MkDocs documentation
│   ├── mkdocs.yml
│   └── docs/
│       ├── index.md
│       ├── installation.md
│       ├── quickstart.md
│       ├── sources/           # Per-source API docs (36 pages)
│       ├── api/               # API reference
│       │   ├── base.md
│       │   ├── registry.md
│       │   ├── cli.md
│       │   └── utils.md
│       └── examples/          # Jupyter notebooks
├── mkdocs.yml                 # Docs config
├── .readthedocs.yaml          # ReadTheDocs config
├── pyproject.toml             # Package configuration
└── README.md

🌐 Available Datasets

Global 🌍

North America 🇺🇸🇨🇦🇲🇽

South America 🌎

*Note: DATASUS access requires pip install epidatasets[brazil] (installs PySUS).

Europe 🇪🇺

Africa 🌍

Asia 🌏

Oceania 🇦🇺🇳🇿

💻 CLI Usage

The epidatasets CLI provides quick access from the terminal (requires pip install epidatasets[cli]):

# List all available data sources
epidatasets sources

# Show detailed info about a source
epidatasets info who

# List countries covered by a source
epidatasets countries paho

News-feed aggregator subcommands:

epidatasets news feeds                                  # list feeds
epidatasets news fetch --days 30 -o items.csv           # fetch + tag
epidatasets news fetch --query "dengue when:30d" -o dengue.csv
epidatasets news plot items.csv --timeline -o timeline.png
epidatasets news map items.csv -o news_map.html         # interactive HTML
epidatasets news animate items.csv -o news_map.gif      # cumulative GIF

💡 Usage Examples

Example 0: Public-health news feeds (early-warning signals)

from epidatasets.newsfeeds import NewsAggregator

agg = NewsAggregator()                       # feeds cached on disk (TTL 6h)
items = agg.fetch(days=30)                   # WHO, PAHO, CDC, CIDRAP, ...
locations = agg.geotag(items)                # country/city mentions -> iso3, lat, lon

agg.plot_timeline(items, by="disease_tags", out="timeline.png")
agg.plot_map(locations, out="news_map.html")
agg.animate_map(locations, out="news_map.gif")

Requires pip install "epidatasets[news,geo]". See the news-feed aggregator docs for the full feed inventory and licensing notes.

Example 1: WHO Global Health Data

from epidatasets import get_source

who = get_source("who")

# Get malaria incidence data
data = who.get_indicator(
    indicator="MALARIA_EST_INCIDENCE",
    years=[2020, 2021, 2022],
    countries=["BRA", "IND", "NGA"]
)
print(data.head())

Example 2: PAHO Pan-American Health Data

from epidatasets import get_source

paho = get_source("paho")

# List member countries
countries = paho.list_countries()
print(f"Total countries: {len(countries)}")

# Get immunization coverage
coverage = paho.get_immunization_coverage(
    vaccines=['DTP3', 'MCV1'],
    subregion='Southern Cone',
    years=[2020, 2021, 2022]
)

# Compare health indicators
comparison = paho.compare_countries(
    indicator='LIFE_EXPECTANCY',
    countries=['BRA', 'MEX', 'ARG', 'COL'],
    years=[2019, 2020, 2021]
)

Example 3: Eurostat EU Health Statistics

from epidatasets import get_source

eurostat = get_source("eurostat")

# Healthcare expenditure
expenditure = eurostat.get_healthcare_expenditure(
    countries=['DEU', 'FRA', 'ITA'],
    years=list(range(2015, 2024))
)

# Mortality data by cause
mortality = eurostat.get_mortality_data(
    cause_code='COVID-19',
    countries=['DEU', 'FRA', 'ITA'],
    years=[2020, 2021, 2022]
)

# Life expectancy comparison
life_exp = eurostat.get_life_expectancy(
    countries=['DEU', 'FRA', 'ITA', 'ESP'],
    years=[2019, 2020, 2021]
)

Example 4: Our World in Data

from epidatasets import get_source

owid = get_source("owid")

# COVID-19 data for specific countries
covid = owid.get_covid_data(
    countries=['BRA', 'USA', 'IND'],
    metrics=['cases', 'deaths', 'hospitalizations'],
    start_date='2021-01-01',
    end_date='2021-12-31'
)

# Excess mortality estimates
excess = owid.get_excess_mortality(
    countries=['GBR', 'ITA', 'USA'],
    start_date='2020-03-01'
)

# Global summary
summary = owid.get_global_summary()

Example 5: Brazil DATASUS via PySUS

from epidatasets import get_source

datasus = get_source("datasus")

# Access Brazilian notifiable disease data
dengue = datasus.download(
    disease="Dengue",
    years=[2022, 2023],
    states=["RJ", "SP", "MG"]
)

Example 6: Africa CDC Data

from epidatasets import get_source

africa_cdc = get_source("africa_cdc")

# List all 55 African Union member states
countries = africa_cdc.list_countries()

# Get disease outbreaks
ebola = africa_cdc.get_disease_outbreaks(
    disease='EBOLA',
    countries=['CD', 'UG', 'GN']
)

# Vaccination coverage
vax = africa_cdc.get_vaccination_coverage(
    countries=['NG', 'ET', 'ZA'],
    vaccines=['COVID-19', 'Measles']
)

Example 7: RKI Germany Surveillance

from epidatasets import get_source

rki = get_source("rki")

# COVID-19 nowcasting with R estimates
nowcast = rki.get_covid_nowcast(
    date_range=('2022-01-01', '2022-06-30')
)

# Influenza surveillance
flu = rki.get_influenza_data(seasons=['2022/23', '2023/24'])

Example 8: Multi-source Comparison

from epidatasets import get_source, list_sources

# See all available sources
print(list_sources().keys())

# Compare data across sources
who = get_source("who")
owid = get_source("owid")

who_malaria = who.get_indicator(
    indicator="MALARIA_EST_INCIDENCE",
    years=[2022],
    countries=["BRA"]
)

owid_covid = owid.get_covid_data(
    countries=["BRA"],
    metrics=["cases", "deaths"],
    start_date='2022-01-01',
    end_date='2022-12-31'
)

Example 9: disease.sh Global COVID-19 & Influenza

from epidatasets.sources.disease_sh import DiseaseShAccessor, DiseaseShAPIError

ds = DiseaseShAccessor()  # optional: cache_ttl_hours=24 for influenza

# Global COVID-19 totals
totals = ds.get_global_totals()

# Per-country current data (name, ISO-2 or ISO-3)
brazil = ds.get_country_data("BRA")

# Historical time series with date filtering
hist = ds.get_historical(
    country=["USA", "BRA"],
    lastdays=90,
    start_date="2023-01-01",
    end_date="2023-03-01",
)

# Vaccine coverage and US CDC influenza surveillance
vax = ds.get_vaccine_coverage(country="USA", lastdays=60)
ili = ds.get_influenza_ilinet()

# Clear errors: invalid country -> ValueError, API failures -> DiseaseShAPIError
try:
    ds.get_country_data("Atlantis")
except ValueError as e:
    print(e)

Example 10: NHS England Prescribing (OpenPrescribing + EPD)

from epidatasets.sources.nhs_prescribing import NHSPrescribingAccessor

nhs = NHSPrescribingAccessor()  # filesystem cache, 7-day TTL (monthly releases)

# National monthly spending on antibacterials (BNF 5.1)
spending = nhs.get_spending(bnf_code="5.1")

# Spending by CCG/SICBL for a single month
ccg = nhs.get_spending_by_ccg(bnf_code="5.1", year_month="2024-01")

# Practice-level spending within a CCG
practice = nhs.get_spending_by_practice(bnf_code="5.1", ccg_code="15N", year_month="2024-01")

# Prescribing quality measures (e.g. antibiotic stewardship)
measures = nhs.get_measures(measure="ktt9_antibiotics", org_type="practice", org_code="L81001")

# Practice locations for spatial analysis (lat/lon from GeoJSON)
practices = nhs.get_practice_locations(ccg_code="15N")

# NHSBSA English Prescribing Dataset (EPD) bulk CSVs — practice-level,
# SNOMED-coded, ~600M rows/year. Chunked + BNF-filtered reading:
chunks = nhs.get_epd_month(2024, 1, bnf_section="5.1", chunksize=500_000)
for chunk in chunks:  # lazy evaluation for large files
    ...

# Monthly time-series for a single chemical (e.g. amoxicillin)
ts = nhs.get_epd_timeseries("0501013B0", "2024-01", "2024-06")

Example 11: Brazilian SRAG line-level surveillance (Sivep-Gripe)

from epidatasets.sources.srag_opendata_sus import SRAGOpenDataSUSAccessor

srag = SRAGOpenDataSUSAccessor()

# Years available on the OpenData SUS portal
srag.available_years()  # [2019, 2020, ..., 2026]

# One year of line-level records (chunked download + lazy reading;
# 0.5-3 GB per year). All filters are composable:
df = srag.get_year(
    2024,
    state="RJ",                  # residence state
    municipality_code=3304557,   # IBGE municipality (6 or 7 digits)
    week_start=1, week_end=26,   # epidemiological weeks
    pathogen="SARS-CoV-2",       # or "Influenza", "VSR"/"RSV", ...
    outcome="death",             # or "cure"
    age_min=60,
    columns=["DT_NOTIFIC", "SEM_NOT", "SG_UF", "EVOLUCAO"],  # narrow early
)

# Memory-bounded chunk iteration for very large years
for chunk in srag.get_year(2022, chunksize=500_000):
    ...

# Weekly counts by geography and pathogen across years
ts = srag.get_weekly_counts(
    years=[2022, 2023, 2024],
    groupby="state",
    pathogens=["SARS-CoV-2", "Influenza", "VSR"],
)

# IBGE municipality helpers
srag.municipality_name(3304557)          # 'Rio de Janeiro'
srag.municipality_code("São Paulo")      # '3550308'

Example 12: Global dengue surveillance (OpenDengue)

Citation required (CC BY 4.0): Clarke J, Lim A, Gupte P, Pigott DM, van Panhuis WG, Brady OJ. A global dataset of publicly available dengue case count data. Sci Data 11, 296 (2024). https://doi.org/10.1038/s41597-024-03120-7

from epidatasets.sources.opendengue import OpenDengueAccessor

od = OpenDengueAccessor()
print(od.citation())  # mandatory citation for CC BY 4.0

# Global extracts: "national", "highest_spatial", "highest_temporal"
global_national = od.get_global(level="national")

# Countries covered (name + ISO3)
od.list_countries()

# Country series with fuzzy name matching and ISO3 codes
bra = od.get_country("Brasil")                      # fuzzy -> BRAZIL
bra = od.get_country(
    "BRA",
    start_date="2015-01-01",
    end_date="2024-12-31",
    admin_level=1,        # state-level series only
)

# One location's time-series
rj = od.get_timeseries("Brazil", admin1="Rio de Janeiro")

# Underlying data-source series per country
od.get_sources(country="India")

📊 Available Sources

❓ FAQ

What is epidatasets?

A Python library providing a unified interface to 41 epidemiological data sources worldwide, installable via pip install epidatasets.

Do I need to install all optional dependencies?

No. The base install covers most sources. Only install extras for sources that need them (e.g., pip install epidatasets[who] for WHO GHO data, pip install epidatasets[brazil] for DATASUS).

How do I discover available sources?

from epidatasets import list_sources
print(list_sources())

Or from the CLI: epidatasets sources

Are all dataset accessors fully implemented?

Most accessors provide working data retrieval. Some are structured placeholders for sources that require registration or have limited public APIs. Check the documentation for each source's status.

Can I contribute a new data source?

Yes! Sources are registered via entry_points in pyproject.toml. See CONTRIBUTING.md for guidelines on adding new accessors.

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for detailed guidelines.

Quick Links for Contributors

• 📋 Contributing Guide - How to get started
• 🐛 Report a Bug
• 💡 Request a Feature
• 🌍 Request a Data Source
• 💬 GitHub Discussions - Ask questions, share ideas

Priority Contributions

1. New data source accessors - Especially from underrepresented regions
2. Example notebooks - Jupyter notebooks demonstrating data analysis
3. Documentation - Translations, improvements, and API docs
4. Bug fixes - Check the issue tracker

Badges for Contributors

📚 Related Projects

📊 Statistics

• Data sources: 39 registered (via plugin registry)
• Countries covered: 100+
• Optional extras: 15 (who, brazil, eurostat, earthengine, climate, geo, viz, genomics, cli, worldbank, search, scraping, gisaid, pdf, news)
• Example notebooks: 47
• Documentation: epidatasets.readthedocs.io

📚 Citation

If you use this package in your research, please cite:

@misc{fccoelho_epidatasets,
  author = {Coelho, Flávio Codeço},
  title = {Epidatasets: Python Access to Epidemiological Datasets Worldwide},
  year = {2026},
  publisher = {GitHub},
  journal = {GitHub Repository},
  howpublished = {\url{https://github.com/fccoelho/epidemiological-datasets}}
}

For PySUS:

@software{pysus,
  author = {AlertaDengue Team},
  title = {PySUS: Tools for Brazilian Public Health Data},
  url = {https://github.com/AlertaDengue/PySUS}
}

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

💜 Sponsor

This project is sponsored by

Kwar-AI — Intelligence for Epidemiology

AI-powered solutions for disease surveillance and outbreak prediction

🙏 Acknowledgments

• PySUS Contributors - For making Brazilian health data accessible
• WHO - For maintaining the Global Health Observatory
• All data providers who make epidemiological data openly accessible
• Global public health community

📞 Contact

• Author: Flávio Codeço Coelho (@fccoelho)
• Website: https://fccoelho.github.io/
• Documentation: https://epidatasets.readthedocs.io

Made with ❤️ for the epidemiological research community

🐛 Report Bug • 💡 Request Feature • 💬 Discussions

Disclaimer: This repository is a community effort to catalog open data sources. Please always refer to the original data providers for official statistics and verify data usage terms. The maintainers are not responsible for data quality or availability.