Reddit financial image post sentiment dataset
Published: 24 October 2022| Version 3 | DOI: 10.17632/b6ns6d8xv3.3
Contributors:
Jonathan Pfahler, , julian wustl, Description
This dataset consists of sentiment information extracted from image and text data of financial subreddit posts. Posts from different financial subreddits are processed using AI tools to create sentiment variables. The data consists of time series data and is fully anonymized. Financial tickers are included to allow financial forecasting.
Files
Steps to reproduce
1) Use API calls to collect posts from the considered financial subreddits for the considered time frame. 2) Train an artifical neural network to classify the images into four groups. 3) Apply sentiment extraction methods to extract sentiment from the posts.
Institutions
Universitat Augsburg
Categories
Finance, Social Media, Machine Learning