Dataset
sp500-edgar-10k
10-K filings for S&P 500 companies retrieved from EDGAR, stored as Parquet with the filing text attached (MIT). A focused corpus for the large-cap annual disclosure cycle: risk factors, MD&A and segment detail for every S&P 500 name.
Data preview
Dataset card
sp500-edgar-10k is a Hugging Face dataset of SEC Form 10-K annual reports for historical S&P 500 constituents filed between 2010 and 2022, paired with forward stock returns. The corpus is stored in Parquet format and is sized in the 1K–10K row range.
What this dataset contains
The dataset bundles parsed text from each 10-K filing with identifier fields such as cik, sic, company, and the filing date, alongside a ret column and 1_day_return, 3_day_return, 5_day_return, and 10_day_return measures. Returns are described as n-day future returns computed manually from external price data, and constituent selection is based on historical S&P 500 membership to mitigate survivorship bias. The text portion preserves the standard 10-K item structure, enabling item-level analysis of business description, risk factors, and financial discussion.
Structure and fields
The schema is tabular with one row per filing. Identifier and metadata columns include cik, sic, company, and a timestamp date field. Text columns follow the 10-K item numbering, covering item_1 through item_15 (including item_1A risk factors, item_7 management's discussion, item_7A quantitative disclosures, and item_8 financial statements), allowing analysts to isolate specific sections. Numeric columns include mkt_cap and the ret, 1_day_return, 3_day_return, 5_day_return, and 10_day_return series for downstream modeling. The single provided split, train, contains 6,282 rows.
How analysts use it
Because the dataset pairs raw filing text with forward return labels, it is suited to training NLP models that link disclosure language to subsequent stock performance, such as risk-factor or MD&A sentiment scoring tied to 5-day or 10-day returns. The cik, sic, and company fields support sector and identifier joins against other fundamental or pricing sources. The historical S&P 500 construction also makes it useful for backtesting event-driven or filing-based strategies across a 2010–2022 window.
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.