Dataset
earnings-call-transcripts
An English-language NLP corpus of cleaned earnings call transcripts gathered from public investor-relations pages, sized between 10,000 and 100,000 documents for financial analysis and LLM work.
Data preview
Dataset card
Rogersurf's earnings-call-transcripts dataset is a cleaned, parquet-formatted corpus of publicly available earnings call transcripts prepared for financial NLP research. It bundles raw transcripts with light metadata and a few derived text fields suitable for downstream language-model tasks.
What this dataset contains
The dataset comprises 9,069 rows of earnings call transcripts collected from publicly accessible transcript pages and processed through a custom ETL pipeline that includes extraction, cleaning and normalization, SQLite persistence, and Pandas-based validation. Each record carries identifying metadata such as ticker, company, quarter, earnings_year, call_date, and title, alongside the cleaned transcript text and a source_url captured at scrape time. The publisher flags the corpus as intended for research and educational use only and reminds users to respect the terms of the original transcript providers.
Structure and fields
Data is delivered as a single parquet file with one training split. The schema mixes numeric and large_string columns: id (int64), ticker, company, quarter, earnings_year (int64), call_date, title, transcript, source_url, scraped_at, and several derived text fields including article_intro, takeaways, summary, glossary, and transcript_clean. This layout preserves the raw transcript alongside lightly enriched summaries and a glossary, letting users choose between the original text and cleaner variants for modeling.
How analysts use it
Given its size and transcript-focused content, the corpus is positioned by the publisher for retrieval-augmented generation, earnings sentiment analysis, financial Q&A systems, market intelligence research, and temporal financial analysis. Quantitative investors could plausibly prototype NLP signal work on guidance language or management tone, while research teams may benchmark summarization and entity extraction models on the included takeaways, summary, and glossary fields. Any production use should be paired with a compliance review of the underlying transcript sources.
Documentation & references 1
Related datasets
Browse moreParsed eligibility criteria from 13,229 ClinicalTrials.gov studies representing the candidate pool gathered by two first-stage retrievers during TREC Clinical Trials 2021–2023, formatted as typed entity-relation graphs for reranking models.
The clinical-trials-patient-graphs dataset is a small tabular collection of patient-level clinical records distributed in Parquet format by the 2001jdev publisher. It organizes entities and relations extracted from trial data into structured rows spanning 2021 through 2023.
This dataset is a parsed snapshot of ClinicalTrials.gov records, distributed in Parquet format on the Hugging Face Hub under an unspecified license. It contains roughly 52,000 trial entries drawn from public registry filings.
clinical-trials-trec-qrels is a tabular relevance-judgment file distributed by the 2001jdev user on the Hugging Face Hub. It maps clinical-trial topic identifiers to NCT registry IDs with graded relevance scores.
The clinical-trials-trec-topics dataset is a small tabular collection of clinical-trial topic records distributed in Parquet format on the Hugging Face Hub. It appears to be derived from TREC clinical-trial retrieval topics, with rows partitioned by year.
SEC EDGAR 10-K & 10-Q financial statements for 2,600+ NASDAQ companies.
Discussion
Questions or observations about this dataset? Join the conversation on X.
No comments yet. Start the conversation on X.