Dataset
earnings-call-data
A static tabular extension of the Bose345/sp500_earnings_transcripts collection covering the same 2005 to 2025 calendar of S&P 500 earnings events, where each row represents a single company-quarter call identified by a stable episode_id and bundles the full transcript along with related SEC press materials and pre-event inputs suitable for supervised learning or reinforcement-style experimentation.
Data preview
Dataset card
The Amaanaush earnings-call-data dataset bundles merged S&P 500 earnings episodes from 2005 to 2025, combining transcripts with price and SEC EDGAR press-release text in parquet form. It is distributed under an MIT license as declared in the dataset card.
What this dataset contains
The release includes a primary episodes.parquet of roughly 33k rows and an episodes_press_release_8k.parquet of about 16k rows, covering hundreds of tickers across a 2005–2025 span. Rows combine upstream Hugging Face transcripts, Yahoo Finance price pulls, and SEC EDGAR 8-K press text, with optional SEC XBRL companyfacts joins available for episodes from 2009 onward. The Hugging Face splits view reports 183,414 train rows in total.
Structure and fields
Each episode is keyed by episode_id and tagged with symbol, company_name, company_id, year, quarter, date, earnings_date, and sector. Text fields hold the earnings_transcript along with press_release_8k_body, press_release_ex991, press_release_ex992, and press_release_sources. Engineered features include guidance_mentioned and beat_mentioned boolean labels plus price_momentum_30d, price_momentum_90d, pct_from_52w_high_pt, and avg_volume_20d, alongside d_minus_1 and d_plus_1 open/high/low/close/volume fields sourced from Yahoo Finance.
How analysts use it
Quantitative equity teams can join the boolean labels such as guidance_mentioned and beat_mentioned with the d_plus_1_close and d_minus_1_close fields to study post-earnings drift and surprise reactions across the symbol and sector columns. NLP and event-driven workflows can treat earnings_transcript and press_release_ex991 as document inputs for classification, summarization, or retrieval tasks, optionally enriched with price_momentum_30d and pct_from_52w_high_pt as conditioning features. Users redistributing derived outputs remain responsible for upstream terms covering the Hugging Face transcripts, yfinance, and SEC data.
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.