Eco Connectivity Workflow

A modular, reproducible workflow for modeling ecological connectivity, integrating database management, machine learning, GIS visualization, and data ingestion.

Overview

The initiative is structured as a modular GitHub stack, allowing multiple teams to work in parallel while keeping the workflow fully reproducible.

Modular Components

Module	Purpose	Repository
Database	PostgreSQL/PostGIS schema and workflow views	db-schema
Machine Learning	Python scripts for resistance modeling and connectivity analysis	ml-models
GIS Projects	QGIS projects, styles, and figure templates	qgis-projects
Data Management	Ingests environmental rasters and camera trap data	data-management
Reproducibility Logs	Tracks scenarios, git commits, and outputs for traceability	reproducibility-logs

---

Workflow Diagram

flowchart TD
    A[Data Team] -->|Ingests detections & rasters| B[Database]
    B -->|Stores schema & views| C[ML Team]
    C -->|Predicts resistance & connectivity| B
    C --> D[GIS Team]
    D -->|Visualizes & exports maps| E[Figures & Publications]
    A --> F[Reproducibility Logs]
    B --> F
    C --> F

Loading

How It Works

1. Data Team – Prepares and ingests environmental rasters and camera trap data into the database.
2. Database (DB) Team – Maintains PostGIS schema, tables, and workflow views.
3. ML Team – Runs resistance and connectivity models, storing outputs back in the database.
4. GIS Team – Connects to the database in QGIS, visualizes results, and exports maps.
5. Reproducibility Logs – Tracks all actions, scenario IDs, and git commits for full traceability.

Getting Started

Technical Stack

• PostgreSQL + PostGIS (or Docker setup)
• Python ≥ 3.9
• QGIS ≥ 3.28 (for project visualization)
• Git

Steps

1. Clone the module repositories:

git clone https://github.com/eco-connectivity-lab/db-schema.git
git clone https://github.com/eco-connectivity-lab/ml-models.git
git clone https://github.com/eco-connectivity-lab/qgis-projects.git
git clone https://github.com/eco-connectivity-lab/data-management.git
git clone https://github.com/eco-connectivity-lab/reproducibility-logs.git

2. Set up the database:

cd db-schema
./setup_db.sh

3. Run ML scripts on sample/test data:

cd ml-models
python scripts/extract_features.py
python scripts/resistance_model.py
python scripts/predict_resistance.py

4. Open QGIS projects for visualization:

cd qgis-projects/projects
qgis eco_connectivity.qgz

5. Track processing with reproducibility logs:

cd reproducibility-logs
python scripts/update_logs.py

Upcoming Enhancements

• Docker & Docker Compose: Launch a full environment (PostGIS + Python + TimescaleDB) with one command.
• CI/CD: GitHub Actions can test SQL views, Python scripts, and QGIS project validity.
• Cloud Integration: Large raster datasets stored as Cloud-Optimized GeoTIFFs (COGs) on S3 or GCS, referenced in the database.

License

GNU GPL v3 license – freely explore and adapt this workflow.