The Atlantic Archive (1857–1930)

Complete pre-1930 archive of The Atlantic, America's most enduring literary and cultural magazine, founded in 1857 by Emerson, Longfellow, Holmes, and other literary giants. **240,011 rows** of clean, structured text spanning from the first volume through 1930.

**What this data does for your model:**

- Your model learns authentic 19th‑century literary voice from Emerson, Longfellow, Holmes, and other founding contributors.
- Your model retrieves actual Gilded Age social commentary on architecture, domestic life, politics, and cultural norms.
- Your model trains on serialized fiction, fairy tales, and literary prose as originally published.
- Your model understands the evolution of American magazine writing from moral essays to cultural criticism.

**Includes:** Volume 1, Number 1 (1857) through 1930.<br/>**Total rows:** 240,011

**What's inside:**

- Literary fiction and poetry from America's greatest writers
- Cultural criticism and social commentary
- Essays on politics, philosophy, and science
- Domestic architecture and design criticism
- Fairy tales and serialized fiction

**Perfect for:**

- LLM fine-tuning on 19th-century American literature
- Cultural history and literary criticism
- Digital humanities and periodical studies
- Journalism and media history research

**Format:** Snowflake-native JSONL with columns: ISSUE, TITLE, AUTHOR, TYPE, TEXT. Fully cleaned, bias-audited, and ready for AI training.

*From Emerson to the Great Depression, the magazine that shaped American letters, now ready for AI.*

<p><br/></p>

**Sample Cortex Agents Prompts:**

<p><br/></p>

*Analyze the historical evolution of American political discourse, literary criticism, and socio-cultural commentary as documented in this long-running archive to support the training of domain-specific AI models focused on the history of intellectual thought.

*Extract longitudinal trends in public opinion, policy advocacy, and the shifting societal debates captured within these pages to assist in the creation of RAG applications for historical research and comparative sociological analysis.

*Evaluate the expert discourse on domestic issues, international affairs, and the progression of American journalistic standards within this dataset to provide a foundational baseline for NLP tasks concerning the history of the 19th and 20th-century media landscape.

<p><br/></p>

原始信息汇总

数据集概述：The Atlantic Archive (1857–1930)

基本信息

• 数据集名称：The Atlantic Archive (1857–1930)
• 提供方：Devin Media Corp.
• 数据规模：240,011 行，涵盖从首卷到1930年的完整、结构化文本。
• 数据格式：Snowflake-native JSONL，包含字段：ISSUE、TITLE、AUTHOR、TYPE、TEXT。
• 数据质量：经过专业清洗、偏差审核，可直接用于AI训练。
• 更新频率：每年更新一次。
• 交付方式：安全共享（Secure share）。

数据内容

包含《大西洋月刊》自1857年创刊至1930年的完整档案，涵盖以下内容类型：

• 文学小说与诗歌（来自美国最伟大作家）
• 文化批评与社会评论
• 政治、哲学与科学论文
• 国内建筑与设计批评
• 童话与连载小说

适用场景

• LLM微调：用于19世纪美国文学领域的大语言模型微调。
• 文化历史与文学批评：支持文化史和文学批评研究。
• 数字人文学科：适用于期刊研究和数字人文学分析。
• 新闻与媒体历史研究：用于媒体史和新闻史研究。

商业需求匹配

• 机器学习：用于训练、微调及部署机器学习模型，适合领域特定LLM微调、文学术语提取和文化自然语言处理。
• 真实世界数据（RWD）：提供1857–1930年间美国知识生活的历史记录，用于研究分析。
• 生命科学商业化：支持人文学科研究，记录美国文学从战前时期到1930年的演变。

使用示例（SQL查询）

• 查看元数据文档：SELECT TITLE, TEXT FROM ATLANTIC_CORPUS WHERE TYPE = metadata LIMIT 5;
• 搜索文学内容：SELECT ISSUE, TITLE, AUTHOR FROM ATLANTIC_CORPUS WHERE TYPE = article AND TEXT ILIKE %poetry% OR TEXT ILIKE %fiction% OR TEXT ILIKE %essay% LIMIT 10;
• 按类型统计行数：SELECT TYPE, COUNT(*) FROM ATLANTIC_CORPUS GROUP BY TYPE;
• 查看1857年首卷：SELECT TITLE, ISSUE, AUTHOR FROM ATLANTIC_CORPUS WHERE TYPE = article AND ISSUE LIKE 1857% LIMIT 10;
• 搜索文化批评：SELECT TITLE, AUTHOR, ISSUE FROM ATLANTIC_CORPUS WHERE TYPE = article AND TEXT ILIKE %architecture% OR TEXT ILIKE %domestic% OR TEXT ILIKE %society% LIMIT 10;

分类标签

• AI & ML
• Life Sciences Commercialization
• Machine Learning
• Real World Data (RWD)

联系方式

• 销售与支持：[e-mail]