Skip to content

SPoRC

Structured Podcast Open Research Corpus is a Python package for working with the SPoRC dataset. It covers 228,099 podcasts and 1,124,058 episodes of transcripts, speaker turns, and acoustic features.

from sporc import SPORCDataset

sporc = SPORCDataset(subset=["The NPR Politics Podcast"])
podcast = sporc.search_podcast("The NPR Politics Podcast")

for episode in podcast.episodes:
    print(f"{episode.title} — {episode.duration_minutes:.0f} min")

Start here

  • Installation. Accept the dataset terms, authenticate with hf auth login, and pip install sporc.
  • Quick start. Load a real podcast and read its episodes and turns in a dozen lines.
  • Working with the data. The corpus is ~57 GB and partitioned by podcast; learn to fetch only the slice you need.
  • Tutorials. Eight notebooks built around real research questions, from corpus cartography to sociophonetics.

What's in the corpus

  • The corpus is a two-month snapshot. Every episode was published between 1 May and 30 June 2020. It is not a longitudinal archive, but the window straddles a sharp, dateable event, which makes before/after designs unusually clean.
  • Every episode has a transcript, but only ~65% have turns. 731,101 of 1,124,058 episodes were diarized into speaker turns; the rest have a transcript but no turns. Always gate on episode.has_turn_data.
  • The acoustics are thin and per-turn. Each turn has six averaged numbers (four MFCCs, one F0, one F1) and no word-level timing. The phonetics module re-derives alignment from source audio when you need it.

The dataset itself (schema, columns, layout, terms of use) is documented on the dataset card. These pages document the Python package.

How the docs are organised

  • Getting started: install, a working quick start, and the data-transfer model you need to understand before pulling anything large.
  • Tutorials: end-to-end notebooks.
  • Guides: task-focused walkthroughs for searching, data loading, categories, conversation analysis, sliding windows, phonetics, and an FAQ.
  • API reference: generated from the docstrings, so it always matches the installed version.

Citation

If you use SPoRC in your research, please cite:

@inproceedings{litterer-etal-2025-mapping,
    title = "Mapping the Podcast Ecosystem with the Structured Podcast Research Corpus",
    author = "Litterer, Benjamin Roger and Jurgens, David and Card, Dallas",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.acl-long.1222/",
    doi = "10.18653/v1/2025.acl-long.1222",
    pages = "25132--25154",
}