josemancharo/apptek_callcenter_dialogues_travel_hospitality_no_transcripts

AppTek Call-Center Dialogues — Travel and Hospitality (No Transcripts)

This is a filtered derivative of AppTek Call-Center Dialogues, prepared for a specific use case.

Changes from the source dataset

• Restricted the dataset to the travel and hospitality domains.
• Removed the transcript field (text) entirely.
• Kept the original audio and the domain, gender, and accent metadata.
• Preserved the source dataset's test split.

This dataset has transcripts removed and is domain-restricted for my specific use case.

Download

Install the Hugging Face Hub client:

pip install huggingface_hub

Download the full dataset (about 4.54 GB):

python get_data.py --output-dir data/apptek_callcenter_travel_hospitality

Download only the metadata without audio:

python get_data.py --metadata-only --output-dir data/apptek_callcenter_metadata

The script downloads the test directory while preserving its accent and audio folder structure. Run python get_data.py --help for all options.

Dataset size

Domain	Audio files
Travel	114
Hospitality	108
Total	222

Each conversation has two single-channel audio files, one for each speaker.

Data fields

• file_name: Relative path to the audio file.
• domain: Either travel or hospitality.
• gender: Speaker gender metadata from the source dataset.
• accent: English accent group from the source dataset.

There is no transcript or text field in this derivative.

Source and license

The source dataset was created and shared by AppTek.ai:

• Source dataset: https://huggingface.co/datasets/apptek-com/apptek_callcenter_dialogues
• Source paper: https://arxiv.org/abs/2604.27543
• License: CC BY-SA 4.0

This derivative is shared under the same CC BY-SA 4.0 license. Refer to the source dataset card for collection methods, speaker demographics, risks, limitations, and the full citation.

Intended use

This subset is intended for work that needs travel and hospitality call-center audio without bundled reference transcripts. It is derived from an evaluation dataset; review the source dataset's intended-use and out-of-scope-use guidance before use.