🏗️ Efficient Download and Extraction of Google Buildings V3 Data

Detailed blogpost explaining this GitHub repo and scripts can be read here: https://sola.kau.se/deprimap/2025/07/23/google-v3-download/

This repository provides a clean two-step pipeline to download and extract building footprint data from Google Open Buildings V3, using only the tiles and geometries that intersect your region of interest (ROI).

---

🌍 Why This Workflow?

Google’s V3 building dataset is powerful — but downloading and processing all ~178 GB is unnecessary for most use cases. This repo solves that by:

✅ Downloading only the tiles that intersect your ROI
✅ Processing those tiles in chunks, extracting only relevant buildings
✅ Saving output as clean .gpkg files ready for spatial analysis

---

🧭 Workflow Overview

Step 1: Identify and Download Tiles

• Input: ROI polygons (can be single or multiple disjoint features)
• Output: Only those .geojson.gz tiles that intersect your ROI
• 📎 Notebook: notebook1_download_tiles.ipynb

The example shows Algeria as an example with multiple polygons spread across the country as an ROI

---

Step 2: Filter and Save Buildings

• Input: Downloaded tiles and your ROI
• Output: GeoPackage (.gpkg) with buildings clipped to your ROI
• 📎 Notebook: notebook2_filtered_buildings.ipynb

⚠️ Although the tile files have a .geojson.gz extension, they are actually CSVs with WKT geometries — not true GeoJSON. This notebook handles that for you.

---

🗺️ Example Use Case

You can use this pipeline for:

• Extracting buildings in disjoint polygon boundaries
• Urban change studies (morphometric studies)
• Lightweight local modelling
• Avoiding unnecessary storage/processing overhead

---

🙏 Acknowledgements

• Google Research - https://sites.research.google/gr/open-buildings/
• DEPRIMAP Project- https://sola.kau.se/deprimap/

This code was developed as part of the DEPRIMAP project for large-scale urban deprivation analysis.