Python library and module for training machine learning models on point-based data and large rasters.
Features:
- Sampling features from a raster using point geometries
- Model training with sklearn and AutoML with TPOT
- Model performance metrics
- Prediction on large rasters
See the demo for more details on features and a guide on running point-eo on the command line. See the demo notebook for a demo of the Python API.
git clone https://github.com/mikkoim/point-eo.git
cd point-eoInstall miniconda and Mamba first. Then create the environment:
mamba env create -f environment.yamlActivate the environment
conda activate point-eo
and install the package
pip install -e .Now point-eo should be available on the command line. Running
point-eo --helpshows available commands. You can find documentation of each command by running the command name and --help, for example:
point-eo analysis --help