Historical Financial News Scraping for LLM Training
BACKLOG
Updated: Fri Jun 20 2025 00:00:00 GMT+0000 (Coordinated Universal Time) | Started: Fri Jun 20 2025 00:00:00 GMT+0000 (Coordinated Universal Time)
Concept project to scrape and structure one to two years of historical financial news data for training and evaluating LLM based price estimation models.
Notes
Overview
This project explores whether large language models can act as effective decision makers when exposed to historical financial news.
Description
The idea is to scrape and structure one to two years of dated financial news from viable sources, then use this data to train or prompt language models for price estimation and decision making experiments.
The project has not yet started and presents challenges around data sourcing, licensing, and processing complexity. It remains an exploratory research idea.
Previous
Scary CrowGame Jam project built in Unity where players farm pumpkins by day and survive crow attacks by night.
Next
Susquehanna Algothon FinalsFinalist and winner performances at the Susquehanna Algothon trading competition.