Dataset
patents
A large granted-patent text collection from the USPTO, tens of millions of records in Parquet (Apache 2.0). Broad coverage of claims and descriptions for IP-intensity and technology-trend measurement.
Data preview
Dataset card
The 'patents' dataset on Hugging Face aggregates more than ten million patent records distributed as a single train split of 10,220,496 rows. Published by user dvdmrs09 and tagged with the Apache-2.0 license, it offers a tabular and text resource for large-scale patent analytics.
What this dataset contains
The dataset comprises roughly 10.2 million rows of patent documents stored in Parquet format and made available through the Hugging Face datasets, dask, mlcroissant, and polars libraries. Each row corresponds to a single patent record identified by its publication number, with accompanying text describing the invention. The records are designed to support both bulk processing and individual look-up by patent identifier.
Structure and fields
The schema is minimal and consists of three string-typed Value fields: publication_number, which serves as the unique patent identifier; title, which holds the patent title; and description, which contains a textual description of the invention. All three fields are encoded as plain strings, reflecting the dataset's focus on text-based patent content rather than structured metadata such as dates, assignees, or classification codes.
How analysts use it
Analysts can load the file directly with polars or dask to filter or aggregate patents by title or description text. The publication_number field enables joins with external patent classifications, citation databases, or stock-level identifier mappings to study innovation activity among public companies. It can also serve as a corpus for training or evaluating NLP models that classify technology domains or track emerging inventive trends.
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.