Tick Tick Bloom: Harmful Algal Bloom Detection Challenge

This repo comes after author's participation in NASA's machine learning competition for cyanobacterial algal bloom severity classification.

General

• Problem statement: use satellite imagery to detect and classify the severity of cyanobacteria blooms in small, inland water bodies.
• Type: Ordinal regression
• Host: NASA
• Platform: Drivendata
• Competition link: https://www.drivendata.org/competitions/143/tick-tick-bloom/
• Placement: Top 1% (5/1377)
• User Name: Ouranos

DownLoad Raw data Notebooks

Clima
Geomorphology
Satellites Earth Engine
Satellites Planetary Computer

Make Datasets Notebook

Make Datasets

Training and Inference Notebook

Training and inference pipeline below is a simplified version ranked 6th scoring 0.811 on private LB instead of author's best 5th place.

v42

Citing This Work

This code has been used in the research paper "AI-driven multi-source data fusion for algal bloom severity classification in small inland water bodies: Leveraging Sentinel-2, DEM, and NOAA climate data". If you find this code useful, please consider citing it.
A preprint can be found on arxiv.

BibTeX:

@article{nasios2025ai,
  title={AI-driven multi-source data fusion for algal bloom severity classification in small inland water bodies: Leveraging Sentinel-2, DEM, and NOAA climate data},
  author={Nasios, Ioannis},
  journal={arXiv preprint arXiv:2505.03808},
  year={2025}
}