Geospatial Flood Assessment

This repository contains a comprehensive geospatial analysis pipeline for assessing flood impacts using various datasets and machine learning clustering techniques. The outputs include interactive maps, GeoJSON files, and raster visualizations.

Key Concepts

This project uses advanced geospatial techniques and clustering algorithms to analyze flood impacts effectively. Key concepts include

• Normalized Difference Vegetation Index (NDVI): Measures vegetation health by comparing the difference between near-infrared (vegetation reflects strongly) and red light.
• Normalized Difference Water Index (NDWI): Highlights water bodies by analyzing the difference between green and near-infrared light.
• Normalized Difference Built-up Index (NDBI): Detects built-up areas by analyzing the difference between short-wave infrared and near-infrared light.
• DBSCAN (Density-Based Spatial Clustering of Applications with Noise): A machine learning algorithm that identifies dense regions in the dataset as clusters, making it suitable for identifying flood-affected zones.
• HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise): An extension of DBSCAN that provides more robust clustering by building a hierarchy of density-based clusters and extracting the best possible grouping.

Methodology

The analysis workflow includes:

1. Data Preprocessing: Input data such as satellite imagery and geospatial layers are processed to compute indices like NDVI, NDWI, and NDBI.
2. Clustering Algorithms:
   • DBSCAN: Identifies dense areas impacted by floods.
   • HDBSCAN: Hierarchical clustering for better cluster definitions.
3. Cluster Severity Analysis: Calculates severity metrics for clusters.
4. Visualization:
   • Generates GeoJSON and TIF outputs for mapping.
   • Produces interactive visualizations for detailed insights.

How to Use

1. Clone the repository:

   git clone https://github.com/your-username/geospatial-flood-assessment.git

2. Install dependencies for Jupyter Notebook:

   pip install -r requirements.txt

3. Open the Jupyter Notebook:

   jupyter notebook geospatial-flood-assessment.ipynb

4. Explore the results using the provided TIF, GeoJSON, and HTML files.

Outputs

• Raster Outputs: Visualize changes in NDVI, NDWI, and NDBI, along with flood clusters.
• Interactive Maps: Explore flood impacts interactively via HTML files.
• GeoJSON Files: Use in GIS platforms for further analysis.

Dependencies

• Python 3.8+
• Jupyter Notebook
• GeoPandas
• Rasterio
• Scikit-learn
• HDBSCAN
• QGIS (optional for QGZ files)

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss changes.

Acknowledgments

• Satellite imagery and data sources.
• Open-source libraries and tools used in this analysis.