Public notebooks for DeepCube's Use Case 3 on Fire Hazard short-term Forecasting

This repository holds jupyter notebooks for some demonstrable results of this use case.

Prerequisites

The notebooks are tested to run with the dependencies in the requirements.txt file.

To install them in your system, run pip install -r requirements.txt.

Datacube Access and Plotting

The first notebook 1_UC3_Datacube_Access_and_Plotting.ipynb downloads and opens the public datacube with xarray.

Then, we demonstrate how the datacube structure allows us to easily visualize and perform analytics on our data.

Deep Learning (DL) models inference

This second notebook 2_UC3_DL_models_inference.ipynb loads pretrained Deep Learning models that are used to create Fire Danger maps for Greece, extracting the necessary data from the datacube presented in the previous notebook.

The methods to download the models are presented in detail in our relevant paper available on arXiv, which has been accepted to the workshop on Artificial Intelligence for Humanitarian Assistance and Disaster Response at the 35th Conference on Neural Information Processing Systems (NeurIPS 2021).

Deep Learning (DL) Models Explainability

This third notebook 3_UC3_DL_models_XAI.ipynb performs the explainable Artificial Intelligence (xAI) techniques used for extracting more information about the data. It also visualizes some plots for better understanding the results.