eohackathon-better-ai

Contains material and instructions necessary to recreate the Machine Learning/AI hands-on exercises prepared for the BETTER project's session at the EO Joint Big Data Hackathon https://www.ec-better.eu/pages/h2020-eo-big-data-hackathon

Land cover changes and inter-annual vegetation performance analysis using ML algorithms

Getting the experiment

This experiment is hosted in a software repository.

Use git to clone it:

cd /workspace
git clone https://github.com/ec-better/eohackathon-better-ai
cd better_ai

Configuring the Python conda environment for experiment.ipynb

The file env_dmuk_ml.yml contains the Python conda environment for running the notebooks contained in this folder.

From the shell, run:

conda env create --file=env_dmuk_ml.yml

Once the environment configuration is done, you can activate it:

conda activate env_dmuk_ml

Running the experiment

Open the environment.ipynb notebook and update the kernel to use env_dmuk_ml

Run the experiment by executing each of the cells with shift + Enter.

If asked for the credentials, provide your Ellip username and associated Ellip API key.

Improving the experiment in a development branch

This experiment is under version control and uses the git flow method (see [https://datasift.github.io/gitflow/IntroducingGitFlow.html])

If not done previously, clone the experiment repository:

git clone https://github.com/ec-better/eohackathon-better-ai
cd better_ai

Then, checkout the develop branch with:

git checkout develop

At this stage, update the experiment.

When done:

git add -A
git commit -m '<commit message>'
git pull

Finally, do a release with:

ciop-release

Getting the better ai

Configuring the Python conda environment for BetterAI.ipynb

The file env_better_ai.yml contains the Python conda environment for running the notebooks contained in this folder.

From the shell, run:

conda env create --file=env_better_ai.yml

Once the environment configuration is done, you can activate it:

conda activate better_ai

Running the betterAI

Open the BetterAI.ipynb notebook and update the kernel to use better_ai

Run the experiment by executing each of the cells with shift + Enter.

Good luck!