RavenPack Transcripts

The RavenPack Transcripts dataset is derived from a global transcripts archive that dates back to 2003. It covers 20,000 publicly-listed companies in the Americas and EMEA, including 2,000+ in Japan, China, Hong Kong, and South Korea.

Analyst calls around earnings are scheduled quarterly for most publicly listed companies, and similar conference calls cover activities including mergers and acquisitions, or CEO changes. Each disclosure may reveal information that materially impacts company strategy or stock valuation, but reviewing transcripts is time-consuming.

RavenPack measures the sentiment and identifies market-moving events in every transcribed earnings call and industry conference. Users can leverage a proprietary taxonomy of over 7,000 business relevant topics to navigate transcripts across thousands of companies and over 20 years of records.

A knowledge graph with over 12 million entities connecting companies, people, products, facilities, and locations is also available to find unique insights.

RavenPack Research has identified playbooks that use earnings intelligence in transcripts data for alpha generation, from equities to corporate bonds, and macro strategies.

For every entity and event detected in a story, RavenPack provides analytics including:
- Sentiment scores
- Relevance metrics
- Novelty tracking
- Temporal scoring
- Topic tagging

For more information, please visit https://www.ravenpack.com/products/edge/data/transcripts

RavenPack转录本数据集（RavenPack Transcripts dataset）源自可追溯至2003年的全球转录本档案库。该数据集覆盖美洲及欧洲、中东与非洲地区（EMEA）的2万家上市公司，其中日本、中国、中国香港及韩国的上市公司数量超过2000家。

多数上市公司按季度召开业绩分析师电话会议，另有同类电话会议涵盖并购、CEO更迭等企业经营动态。每份披露文件均可能包含对公司战略或股票估值产生重大影响的信息，但人工审阅转录本耗时耗力。

RavenPack可对每一份业绩电话会议及行业会议的转录本开展情感分析，并识别出影响市场的重大事件。用户可借助涵盖7000余个商业相关主题的专有分类体系，在数千家公司、跨度超20年的海量转录本记录中快速检索与浏览。

此外，数据集还提供包含超1200万个实体的知识图谱（knowledge graph），这些实体关联企业、人物、产品、设施及地理位置，助力用户挖掘独特商业洞察。

RavenPack研究团队已开发出多款应用范式，可借助转录本数据中的业绩情报生成超额收益，应用场景覆盖股票、公司债券及宏观策略等领域。

针对文本中检测到的每一个实体与事件，RavenPack均提供以下分析指标与工具：
- 情感评分
- 相关性度量
- 新颖性追踪
- 时序评分
- 主题标注

如需了解更多信息，请访问：https://www.ravenpack.com/products/edge/data/transcripts

搜集汇总

背景与挑战

背景概述

RavenPack Transcripts数据集涵盖2003年以来美洲、EMEA及亚洲地区2万家上市公司的会议记录，提供情感分析和7000+商业主题分类。其特色包括1200万实体知识图谱，支持通过财报电话等文本数据挖掘投资策略洞察。