Skip to content

Dataset · lamini

Financial earnings call transcripts and Q&A pairs

lamini/earnings-calls-qa

Transcripts of quarterly corporate earnings conference calls and English Q&A pairs based on them.

What's inside

Inside are text transcriptions of various companies' conference calls based on financial reporting, along with questions and answers about their financial indicators. The dataset size ranges from 100 thousand to 1 million English records. The data is stored in jsonlines format, where each object contains a single company's call transcript.

What you can do with it

  • Training or fine-tuning QA systems for analyzing corporate financial reporting
  • Training models for fact extraction and summarization of long business transcripts
  • Classifying financial texts and the sentiment of corporate management statements
Tasks
Text classification, Question answering, Text generation
Languages
English
Size
100K–1M rows
Downloads this month
348
Likes
55

How to load it

from datasets import load_dataset

ds = load_dataset("lamini/earnings-calls-qa")