Environmental Monitoring in New Zealand: An In-Depth Exploration of Spatiotemporal Relation

This repository contains the code, data, and experiments related to the paper Environmental Monitoring in New Zealand: An In-Depth Exploration of Spatiotemporal Relation, accepted to the AAAI 2026 Workshop on AI for Environmental Science.

The project is structured to ensure reproducibility, modularity, and comparative analysis across different model families.

---

📁 Repository Structure

├── code/ # Python scripts for each model implementation
├── dataset/ # Datasets used for training and testing
├── forecasting/ # Forecast outputs generated by the models
├── notebooks/ # Jupyter notebooks for analysis and experiments
├── plots/ # Figures and visualizations of the results
├── LICENSE # License file
└── README.md # This file

---

⚙️ Environment Setup

Clone the repository

git clone https://github.com/LabIA-UFBA/TAIAO-forecasting.git
cd repo_taio_aaai

Create a virtual environment (recommended)

python -m venv venv
source venv/bin/activate   # Linux / macOS
venv\Scripts\activate  

Install dependencies

pip install -r requirements.txt

🚀 Running the Models

Each model is implemented in a separate Python file inside the code/ directory. For example:

python code/model_sarima.py
python code/model_lstm.py
python code/model_gru.py
python code/model_gnn.py

📈 Reproducibility notebooks

All experiments can be reproduced using the main notebook:

jupyter notebook notebooks/Cronos-t5-base.ipynb

jupyter notebook notebooks/TIMES_FM.ipynb

Forecast outputs are automatically saved under the forecasting/ folder.

---

📊 Visualizing Results

The notebooks inside notebooks/ allow you to reproduce, compare, and visualize results. Example:

jupyter notebook notebooks/CHEB.ipynb

🧾 License

This project is distributed under the GNU GENERAL PUBLIC LICENSE. See the LICENSE file for details.