🛰️ OSINT-GeoAI-Suite

Applied ML Prototypes for Defence and Strategic Intelligence
Maintained by: Favour 'Nimi' Adebayo
License: MIT / Academic Research

Overview

A cohesive GeoAI-OSINT analytical stack that demonstrates open data triangulation, anomaly detection, and strategic risk assessment—all using reproducible, ethical, and transparent ML. This suite demonstrates how a defence or security analyst can operationalize open-source data (AIS, Sentinel imagery, SIPRI, WB logistics) through machine learning for situational awareness and risk assessment. Each module is self-contained yet interoperable, providing complementary intelligence layers:

Module	Description	Stack	Output
Seance	Detects maritime route anomalies near global choke points	AIS + ML clustering (HDBSCAN + Isolation Forest)	Streamlit map dashboard
Sentinel	Flags construction/logistics buildup along borders using Sentinel-2/SAR imagery and OSINT text	GEE / raster analysis + NLP	Jupyter + technical report
Ecolog	Quantifies supply-chain stress near conflict zones	SIPRI + WB + trade + Prophet forecasting	Time-series dashboard

These prototypes replicate OSINT workflows used in defence intelligence for: - Early warning and pattern-of-life anomaly detection - Cross-domain correlation between physical and economic indicators - Transparent, open-data-based situational awareness - Strategic analysis that supports policy and operational foresight

Analytical Flow

[Seance: Maritime Anomalies] ─┐
                              ├──> [Ecolog SupplyChain Index] ───> Composite Risk Map
[Sentinel: Border Signals]   ─┘

Directory Architecture

osint-geoai-suite/
│
├── README.md                            # master overview of suite
├── requirements.txt                     # core shared dependencies
├── /data/                               # shared lightweight sample datasets (public)
│
├── seasense-maritime-anomaly/
│   ├── README.md
│   ├── data/
│   │   ├── ais_sample.csv
│   ├── notebooks/
│   │   ├── 01_preprocessing.ipynb
│   │   ├── 02_clustering_anomalies.ipynb
│   ├── app/
│   │   ├── dashboard.py                 # Streamlit app
│   ├── src/
│   │   ├── preprocessing.py
│   │   ├── anomaly_model.py
│   │   ├── visualization.py
│   ├── docs/
│   │   ├── analysis_report.md
│   └── environment.yml
│
├── sentinelwatch-border-activity/
│   ├── README.md
│   ├── data/
│   │   ├── sentinel2_sample.tif
│   ├── notebooks/
│   │   ├── 01_change_detection.ipynb
│   │   ├── 02_text_integration.ipynb
│   ├── src/
│   │   ├── image_preprocessing.py
│   │   ├── change_detection.py
│   │   ├── text_signal_extraction.py
│   ├── docs/
│   │   ├── technical_writeup.md
│   └── environment.yml
│
└── riskflow-supplychain-index/
    ├── README.md
    ├── data/
    │   ├── sipri_spend.csv
    │   ├── wb_logistics.csv
    │   ├── port_throughput.csv
    ├── notebooks/
    │   ├── 01_data_integration.ipynb
    │   ├── 02_index_construction.ipynb
    │   ├── 03_forecasting_validation.ipynb
    ├── app/
    │   ├── dashboard.py                 # Plotly/Dash app
    ├── src/
    │   ├── preprocess.py
    │   ├── build_index.py
    │   ├── forecast_model.py
    ├── docs/
    │   ├── riskflow_report.md
    └── environment.yml

Setup

# Clone the suite
git clone https://github.com/<username>/osint-geoai-suite.git
cd osint-geoai-suite

# Install shared dependencies
pip install -r requirements.txt

# Navigate into each repo and install its environment
cd seance-maritime-anomaly
pip install -r env.yml

Data Ethics

All data are public, non-classified, and aligned with responsible AI principles: reproducible, non-intrusive, and strictly analytical.