datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
atc-pilot-speaker-role-classification-dataset
ATC-Pilot Speaker Role Classification Dataset
This dataset is designed for binary text classification in air traffic communication, distinguishing between utterances made by pilots and those made by air traffic controllers (ATC).
Dataset Origin
The data was derived from the UWB ATC Corpus, a publicly available dataset of air traffic control communications provided by the University of West Bohemia. This corpus consists of approximately 20 hours of manually transcribed… See the full description on the dataset page: https://huggingface.co/datasets/jacktol/atc-pilot-speaker-role-classification-dataset.atc-role-classification-synthetic-2procedurally generated synthetic data, script is included in files
atc-role-classification-syntheticGenerated with Claude Opus 4.6
The data covers a wide range of realistic radio phraseology including:
Pilot utterances: check-ins, readbacks, requests (altitude/heading/approach/deviation), position reports, PIREPs, emergencies (mayday/pan-pan), taxi/pushback, go-arounds, and general acknowledgments.
ATC utterances: clearances (takeoff/landing/approach), altitude/heading/speed instructions, traffic advisories, hold instructions, weather broadcasts, vectors, STAR/SID assignments, taxi… See the full description on the dataset page: https://huggingface.co/datasets/TigreGotico/atc-role-classification-synthetic.atc-role-classification-synthetic-3Generated with ChatGPT