environmental-insights 0.2.8

A Python package for democratizing access to ambient air pollution data and predictive analytics.

Environmental Insights

A Python package for democratizing access to ambient air pollution data and predictive analytics.

---

📖 Description

Environmental Insights provides easy-to-use functions to download, process, and analyze ambient air pollution and meteorological data over England.

• Implements supervised machine-learning pipelines to predict hourly pollutant concentrations on a 1 km² grid.
• Supplies both “typical day” aggregates (percentiles) and full hourly model outputs.
• Includes geospatial utilities for mapping, interpolation, and uncertainty analysis.

---

⚙️ Installation

Install from PyPI:

pip install environmental-insights

Or from source:

git clone https://github.com/liamjberrisford/Environmental-Insights.git
cd Environmental-Insights
python -m build
pip install dist/environmental_insights-0.2.1b0-py3-none-any.whl

---

🔗 Dependencies

• Python ≥ 3.10
• geopandas ≥ 1.0.1
• lightgbm
• matplotlib
• overpy
• pandas
• pyarrow
• pyogrio
• requests
• scipy
• shapely
• jupyterlab ≥ 4.4.2
• xarray ≥ 2025.4.0
• netcdf4 ≥ 1.7.2
• scikit-learn ≥ 1.6.1

---

📂 Data Sources

This package downloads and processes two primary CEDA datasets:

1. Synthetic Hourly Air Pollution Prediction Averages for England (SynthHAPPE)
   Berrisford, L. (2025). Synthetic Hourly Air Pollution Prediction Averages for England (SynthHAPPE). NERC EDS Centre for Environmental Data Analysis.
   DOI: 10.5285/4cbd9c53ab07497ba42de5043d1f414b

   Representative “typical day” profiles of NO₂, NO, NOₓ, O₃, PM₁₀, PM₂.₅ and SO₂ on a 1 km² grid, with 5th, 50th & 95th percentiles.

2. Machine Learning for Hourly Air Pollution Prediction in England (ML-HAPPE)
   Berrisford, L. (2025). Machine Learning for Hourly Air Pollution Prediction in England (ML-HAPPE). NERC EDS Centre for Environmental Data Analysis.
   DOI: 10.5285/fc735f9878ed43e293b85f85e40df24d

   Full-year (2018) hourly modelled concentrations of NO₂, NO, NOₓ, O₃, PM₁₀, PM₂.₅ and SO₂ on a 1 km² grid, including 5th, 50th & 95th percentiles and underlying training data.

---

For full examples, see the Jupyter-Book tutorial in book/tutorial_environmental_insights.ipynb.

📚 Documentation

Build and view locally:

jupyter-book build book/

Then open book/_build/html/index.html in your browser.
Highlights:

• API Reference: book/docs/api/environmental_insights/
• Tutorial Notebook: book/tutorial_environmental_insights.ipynb

The documentation is also avaiable via the GitHub Pages Site

---

✅ Testing

Run the full test suite:

pytest

Integration and unit tests are under tests/.

---

🤝 Contributing

Contributions and bug-reports are very welcome! Please see CONTRIBUTING.md for details on:

• Code style
• Pull request process
• Issue reporting

---

📑 Citation

If you use Environmental Insights in your work, please cite:

Berrisford, L. J. (2025). Environmental Insights: Democratizing access to ambient air pollution data and predictive analytics (Version 0.2.1b0) [Software]. GitHub. https://github.com/liamjberrisford/Environmental-Insights

Also cite the underlying datasets:

• Berrisford, L. (2025). SynthHAPPE: Synthetic Hourly Air Pollution Prediction Averages for England. NERC EDS CEDA. DOI: 10.5285/4cbd9c53ab07497ba42de5043d1f414b
• Berrisford, L. (2025). ML-HAPPE: Machine Learning for Hourly Air Pollution Prediction in England. NERC EDS CEDA. DOI: 10.5285/fc735f9878ed43e293b85f85e40df24d

---

📜 License

This project is released under the GPL-3.0-or-later.