Dataset
EDGAR10-Q
Quarterly 10-Q filings parsed from EDGAR into JSON, millions of records in English (MIT). The quarterly companion to the 10-K corpus for tracking sequential changes in operating performance and risk factors.
Data preview
Dataset card
EDGAR10-Q is a large text dataset of annual and quarterly reports scraped from the top 1,500 public companies, designed for the ContextNER task of generating contextual phrases for entities in financial sentences. It contains roughly 1.87 million input-output sentence pairs across train, validation, and test splits.
What this dataset contains
The dataset is built from 10-Q and 10-K filings pulled from the SEC's EDGAR system and is organized as plain text input-output pairs. According to the publisher, it is among the largest financial NLP resources of its kind, with about 1 million sentences, 2.8 million entity mentions, and an average of 35 tokens per sentence. The combined splits total 1,873,328 rows.
Structure and fields
Each row is defined by two string fields: an `input` column containing a sentence that includes a target entity, and an `output` column containing a set of contextual phrases describing that entity. For example, an input referencing a variable interest entity is paired with an output such as "Ownership percentage," and an input mentioning a 2022 Series A Notes tranche is paired with "Stated interest rate." The data is delivered in JSON format and is split into train, validation, and test sets with identical schemas.
How analysts use it
The resource is primarily aimed at NLP researchers building models that can recover or expand the context around financial entities, as described in the associated arXiv paper. For investment workflows, the sentence-level entity and context annotations can support training custom language models to extract financial signals, attributes, or relationships from filings, though downstream use would require additional model development beyond the dataset itself.
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.