Alternative data datasets
Browse by provider, category, marketplace, access, delivery and license
View source information and access options.
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.
French patents published in 2024, split into chunks at the `<p>` XML tag level and saved in Parquet for streaming and distributed processing. The records come from public French patent filings (A1, 2024) and were independently extracted, structured, and segmented using access to PI API and FTP sources.
A cleaned collection of French patent documents published from 1981 through 2026, each row in the Parquet file representing one complete A1 publication parsed from the original XML. It was assembled independently by a contributor with direct access to the public patent source and is optimized for distributed loading.
Raw French patent publications from 2020 through 2026, pulled from original A1 XML records by an independent API/FTP extraction and delivered one-document-per-row in streaming-ready parquet.
A collection of 2,000 Malaysian property transaction records spanning every state offers a detailed look at the national housing market in 2025. Compiled from Brickz, a recognized source for real estate transaction data, it covers location, tenure, property type, median pricing, and transaction volumes.
Earnings-call transcripts from an IBM research project, tens of thousands of records with finance tags, in CC0. A filings-adjacent corpus for call-tone and question-and-answer analysis.
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.
The mintic_linkedin-job-postings dataset aggregates LinkedIn job posting text into a single string column with roughly 124,000 rows. It is published on Hugging Face by user jmparejaz under an unstated license.
Metadata for every EDGAR filing submitted to the U.S. Securities and Exchange Commission from 1994 through December 14, 2024, derived from the quarterly master index files and including CIKs, issuer names, and form types.
patent_abstract_uspto is a text dataset of U.S. patent abstracts distributed under an MIT license and hosted on the Hugging Face Hub by publisher kaustav16. It contains roughly 9.36 million rows of patent identifiers paired with their corresponding abstract text.
sec_filings is a small text dataset of 494 rows that pairs natural-language queries about U.S. securities filings with retrieved document facts and ground-truth answers. It is distributed under an unstated license on a public dataset hub.
The patents-publications-dataset is a large-scale tabular collection of patent publication records distributed in Parquet format. Hosted by publisher labofsahil on the Hugging Face Hub under an unstated license, it is sized in the 100M–1B row range.
The clinical_trials dataset distributes ClinicalTrials.gov registry records in Parquet format for offline analysis. It contains a single training split of roughly 150,000 rows sourced from the public clinical-trial registry.
PatentDesc Dataset Based on data from vl2g licensed under MIT License. Copyright (c) 2024 vl2g.
patent-desc-test is a small text dataset on Hugging Face Datasets that contains short identifier-style strings derived from patent document numbers. The dataset is published by user matthias-ehrlich and is broadly tagged as public records–adjacent text data.
A compact sample set of SEC filings used for the MemGPT agent demos, tens of thousands of text records. A small, clean entry point to filings text rather than a comprehensive corpus.
Patent All Claims EP granted claim sets (indepdendent and dependent claims). One claim per line. Claims from 20210915-20250806 Splits Train: 96% Validation: 2% Test: 2% Usage from datasets import load_dataset ds = load_dataset("mhurhangee/ep-patent-all-claims")
A collection featuring the primary independent claim of United States utility patents issued between 2005 and 2025 is provided. Because the opening claim usually establishes the widest legal boundaries of an invention, it carries particular significance, and each entry includes the patent identifier alongside its claim text.
View source information and access options.
View source information and access options.
google-patents-data-preview is a community-uploaded preview of Google Patents bibliographic records, distributed in Parquet format through the Hugging Face Hub. The single training split contains roughly 340,000 rows covering patent identifiers, classifications, and localized text fields.
Global Job Postings – Multi-ATS is a single dated snapshot of 112,816 job listings aggregated from 23 applicant tracking systems and parsed into a 47-column schema by NextGig-Rocks. It is distributed as Parquet via the Hugging Face datasets library and targets labor-market research, not production job boards.
TUDelft-Electricity-Consumption-1.0 is a high-resolution, open-source time series dataset of household electricity load published on Hugging Face by OpenSynth. It draws on multi-country trials covering thousands of households under different tariff regimes and is distributed in Parquet format.
Clinical-trial protocol documents from Parexel, tens of thousands of studies. Protocol text, covering inclusion criteria, endpoints and timelines, is a leading indicator for enrollment pace and pipeline risk.
Dataset card entry for "AIPD_nlp_granted_claims" with additional details required.
patent-drafting-bench 中文 | English patent-drafting-bench 是一个评估大模型及智能体专利撰写能力的数据集，包含 60 个测试任务。模型需要依据技术交底书，按指定法域和语言撰写权利要求书、说明书及必要附图，最终交付一份 Word（.docx）文档。 评测关注模型能否将交底中的技术方案转化为完整、合理且有依据的专利申请文本，包括权利要求的完整性、清楚性、保护范围和从属布局，说明书的支持与公开，以及是否编造技术内容。每个任务包含通用评分标准与案例特异评分标准，可用于逐项人工审阅或 LLM judge 评分。 目录结构 patent-drafting-bench/ ├── README.md ├── README.en.md ├── data/ │ └── test.jsonl └── asset/ ├── PD-0005_input.docx
专利无效检索评测数据集 面向专利无效证据检索系统的公开 Benchmark，以权利要求文本为检索输入， 以真实无效案件中使用的证据文献为标准答案。 简体中文 · English 下载公开数据集 数据集简介 本数据集用于评估 AI Agent、检索模型和传统专利检索系统发现无效证据文献的能力。 每条样本包含待分析的权利要求、案件最终结果、证据文献，以及证据角色和法律依据等 结构化标注。 公开版本包含 80 条 Benchmark 样本，CN 与 US 各 40 条，覆盖 IPC 的 A–H 八个部。它既可以衡量系统能否找到至少一篇有效证据，也可以衡量系统对核心文献和 全部证据文献的覆盖程度。 这份数据可以用来评估什么 能力 说明 无效证据检索 给定权利要求文本，系统能否从专利文献中找到真实案件使用的无效证据 核心与完整召回 分别评估主要证据 primary_reference 和全部证据文献
PatSnap Patent OAR Bench A benchmark dataset for patent Office Action Response (OAR) tasks. The test split contains 40 cases: 13 CN, 13 EP, and 14 US. It includes 40 Office Action PDF documents totaling 229 pages, structured case materials, and 19 shared Judge-scored criteria per case (760 entries). Dataset: data/test.parquet, 40 cases in Parquet format. Office Action PDFs: assets/, named OAR-<id>
A daily record of trading activity across multiple major global stock market indices, structured for trend analysis and forecasting work in international equity markets.
Patent_FR_US_Merge_Radix_65536 is a Hugging Face dataset by publisher PhysiQuanty that merges French and US patent documents into a single text corpus. It is distributed in Parquet format and is tagged with an Apache-2.0 license, though the README provides no further description.
The ja_patent mirror of llm-jp-corpus-v4, assembled by the LLM-jp Corpus Building Working Group at NII, reproduces the Japanese patent sub-corpus from the larger LLM-jp Corpus v4 release. It is delivered as 621 jsonl.gz files totaling 58.2 GB in compressed form, with each line holding a JSON object that includes a text field and a meta field containing the document identifier, URL, and other provenance information.
The nj-housing-prices-tabular dataset is a small tabular collection of New Jersey residential property listings, published on Hugging Face under an unstated license. It contains roughly 7,000 rows split across train, validation, and test partitions.
clinicaltrials.gov-summary_and_eligibility is a small Parquet mirror of selected ClinicalTrials.gov records, distributed on the Hugging Face Hub under the rjac publisher. It packages study identifiers, recruitment status, titles, brief summaries, and eligibility criteria into a single tidy table of roughly three thousand rows.
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.
Citation This dataset builds on edgar-corpus.
A resource built for forecasting patent approval outcomes, introduced in research employing domain-specific fine-grained claim dependency graphs, with further documentation to be supplied later.
Dataset Card for "filtered_yelp_restaurant_reviews" More Information needed
sec-filings is a Hugging Face dataset published by Sicheng-Chroma that distributes pre-chunked U.S. SEC filings paired with dense vector embeddings. It is distributed in optimized Parquet format and is intended for retrieval and similarity research over regulatory disclosures.
View source information and access options.
Agent Earnings 87 sourced, dated entries on money made with AI agents: agent businesses, marketplace and bounty payouts, x402 pay-per-call sales, sponsor and affiliate income, and other experiments. The set includes flops, one withdrawn figure, one sponsor list price (not revenue), one project count (not a dollar amount), one bounty total (Superteam Earn, $84,137) that the entry marks as claimed b
Earnings Call Transcript Lite is a small public-records dataset of corporate earnings call transcripts paired with short reference summaries. It is distributed as a CSV and is published on the Hugging Face Hub under an unstated license.
tech-job-postings-labeled is a small public dataset of labeled technology job postings published on Hugging Face by SuhaibAtef. It comprises roughly 8,000 records tagged with source labels and is distributed in JSON.
jobpostingsamples is a small sample dataset of US job postings published by talanAI. It is distributed in CSV format and is tagged for use with the Hugging Face datasets, pandas, polars, and mlcroissant libraries.
View source information and access options.
A collection of Parquet files providing one record per Chinese invention patent family granted between 2003 and 2019, capturing a cumulative quality assessment pipeline that includes metrics such as semantic knowledge recombination, strict historical semantic novelty, and IPC-based robustness measures across global patent publications.
10-K annual filings from EDGAR in Parquet, tens of thousands of documents. The standard annual-disclosure corpus for fundamentals, risk-factor and management-discussion analysis.
The linkedin-job-postings dataset, published by xanderios on Hugging Face, contains a single CSV table of U.S. LinkedIn job listings with about 33,000 rows. It is tagged with a MIT license on the platform, though the underlying LinkedIn terms of service place restrictions on redistribution and use.
The 2024 Earnings Call Transcript dataset, published by yeong-hwan on Hugging Face, aggregates quarterly earnings call dialogue for individual stock tickers. It is distributed as a JSON table with 1,904 rows in a single train split.
The Tech Job Postings & Salary Dataset from Zalize Data contains over 449,000 monthly job postings scraped from public applicant tracking systems at US and EU technology companies. It is distributed as parquet files under a CC BY-NC 4.0 license and is part of the DataForge Open Data program.