point-eo

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.

Overview

Installation

git clone https://github.com/mikkoim/point-eo.git
cd point-eo

Install miniconda and Mamba first. Then create the environment:

mamba env create -f environment.yaml

Activate 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 --help

shows available commands. You can find documentation of each command by running the command name and --help, for example:

point-eo analysis --help