cetaceans

banter is a package for creating hierarchical acoustic event classifiers out of multiple call type detectors.

Automatic dolphin counter

Geometric analysis of cetacean communication: Poincaré embeddings, TDA, SPD manifolds, and adversarial decoder robustness

This repo is part of my Masters Project in the course Marine Science and Climate Change at the University of Gibraltar

A deep learning image classification project to predict whale and dolphin species

Repos contains code and data that underpins a novel dolphin machine learning model.

Machine learning project to classify dolphins from drone footage using image and velocity information.

This supplementary material contains the files, scripts and methods that were used to develop the analysis of the research article entitled "Risk Assessment of Potential Tourist-Vessel Collision with cetaceans in the Colombian EEZ".

Preliminary analysis of the questionnaire for face-to-face surveys to the targeted fisher groups engaged in gillnet coastal fisheries.

Open-source interactive globe visualising 25,000+ whale sightings, strandings, acoustic detections and migration routes from 6 species across 5 open scientific databases.

Each script and notebook in the repository is well-documented, with extensive logging for transparency and reproducibility. The repository is designed for researchers, data scientists, and conservationists interested in applying data-driven methods to marine ecology and species management.

Six-stage reproducible pipeline for CNN-based passive acoustic monitoring: audio annotation, segmentation, spectrogram generation, image preprocessing, k-fold training, and cross-domain evaluation. Fully configured via YAML.

Measurement-driven procedural cochlea generator for Blender, with validated GLB assets and turn-count visualizations.

Interactive map of cetacean observations and vessel strike risk in Norwegian and surrounding waters. Conservation screening tool with peer-reviewed collision risk formulas.

Fits a dynamic (multi-season) occupancy model for cetacean vessel-survey sightings in JAGS. Config-driven data cleaning, spatial gridding, environmental covariates, and diagnostics for any species or study area.