Datasetpublic

sanjeevafk/great-speeches-corpus

🎙️ Great Speeches Corpus: Definitive Historical Archives Benchmark (1893–2015) Speaker Breakdown & Catalogue Index Namespace Historical Figure Time Span Speeches Word Count Character Count Primary Domain nehru Pandit Jawaharlal Nehru 1929–1964 5,000 1,315,811 8,636,056 Nation Building, Democracy, NAM mandela Nelson Mandela 1951–2013 1,000 1,141,540 7,083,904 Anti-Apartheid, Reconciliation, Human Rights kalam Dr. A. P. J. Abdul Kalam 1989–2015… See the full description on the dataset page: https://huggingface.co/datasets/sanjeevafk/great-speeches-corpus.

sourceHugging Faceapache-2.0updated 1mo agoView on Hugging Face

0likes47downloads

Dataset Card

🎙️ Great Speeches Corpus: Definitive Historical Archives Benchmark (1893–2015)

![Hugging Face Dataset](https://huggingface.co/datasets/sanjeevafk/great-speeches-corpus) ![License: Apache 2.0](https://opensource.org/licenses/Apache-2.0) ![Dataset Viewer](https://huggingface.co/datasets/sanjeevafk/great-speeches-corpus)

Dataset Description

—Curator: Sanjeevafk Speech Archives Project
—Scope: 9 Iconic World Figures across Politics, Science, Philosophy, Technology, and Civil Rights
—Time Span: 1893 – 2015 (122 Years)
—Total Canonical Speeches: 8,687
—Total Corpus Volume: 5,457,227 words (34,662,197 characters)
—Dataset Hierarchy: Available as a unified dataset (config="all") or partitioned by speaker namespaces (config="kalam", config="nehru", config="mandela", etc.).

Speaker Breakdown & Catalogue Index

Namespace	Historical Figure	Time Span	Speeches	Word Count	Character Count	Primary Domain
`nehru`	Pandit Jawaharlal Nehru	1929–1964	5,000	1,315,811	8,636,056	Nation Building, Democracy, NAM
`mandela`	Nelson Mandela	1951–2013	1,000	1,141,540	7,083,904	Anti-Apartheid, Reconciliation, Human Rights
`kalam`	Dr. A. P. J. Abdul Kalam	1989–2015	912	1,822,794	11,374,971	Vision 2020, Science, Youth Inspiration
`ambedkar`	Dr. B. R. Ambedkar	1916–1956	537	74,465	558,227	Constitution of India, Social Justice, Law
`vivekananda`	Swami Vivekananda	1893–1902	500	123,276	807,740	Four Yogas, Parliament of Religions, Vedanta
`einstein`	Albert Einstein	1909–1955	260	64,815	425,995	Relativity, Nobel Lecture, Nuclear Disarmament
`bose`	Netaji Subhas Chandra Bose	1921–1945	215	51,032	353,909	Azad Hind, INA, Freedom Struggle
`feynman`	Richard P. Feynman	1955–1988	162	845,526	5,300,276	Quantum Electrodynamics, Caltech Lectures
`steve_jobs`	Steve Jobs	1976–2011	101	17,968	121,119	Apple Keynotes, Technology, Design
`all`	Complete Unified Benchmark	1893–2015	8,687	5,457,227	34,662,197	Multidisciplinary Master Corpus

Usage Examples

1. Load Complete Multi-Speaker Corpus

python

from datasets import load_dataset

# Stream the entire 8,687-speech historical benchmark
dataset = load_dataset("sanjeevafk/great-speeches-corpus", "all", split="train")
print(f"Total speeches loaded: {len(dataset)}")

2. Load Specific Historical Speaker Namespaces

python

from datasets import load_dataset

# Load Jawaharlal Nehru (5,000 speeches)
nehru_corpus = load_dataset("sanjeevafk/great-speeches-corpus", "nehru", split="train")

# Load Swami Vivekananda (500 discourses)
vivek_corpus = load_dataset("sanjeevafk/great-speeches-corpus", "vivekananda", split="train")

# Load Albert Einstein (260 scientific lectures)
einstein_corpus = load_dataset("sanjeevafk/great-speeches-corpus", "einstein", split="train")

Schema

—speech_id (string): Canonical identifier (SPEECH-ENTITY-YYYY-NNN)
—canonical_event_id (string): Unique event identifier
—title (string): Title of speech, address, keynote, or lecture
—date (string): Event date (DD.MM.YYYY)
—year (int): Year of presentation
—location (string): City, Country/State
—venue (string): Hall, parliament, university, stadium, or convention center
—period (string): Era classification tag
—event_type (string): presidential_address, keynote, lecture, parliamentary_speech, historic_oration
—speaker (string): Name of the historical speaker
—tldr (string): Multi-sentence summary of core themes and historic significance
—transcript (string): Full verbatim text in markdown format
—source_url (string): Authoritative archive URL
—rights_status (string): Transcript rights status
—sha256 (string): Bit-level SHA-256 integrity hash
—tags (list[string]): Semantic domain and topical tags