Predictive Power of Alternative Data in Financial Markets

Keywords

Abstract

This research investigates the predictive power of alternative data sources in forecasting stock returns. Specifically, we analyze Google Trends search volume data, Glassdoor employee reviews, and consumer sentiment indices to predict future stock performance. Using panel regression analysis and correlation studies across a diverse universe of equities, we find statistically significant predictive relationships between these alternative data signals and subsequent stock returns. Google Trends data shows strong correlation with retail investor interest and momentum effects, while Glassdoor ratings provide early signals of corporate health and employee satisfaction that precede fundamental changes. The study demonstrates that incorporating alternative data into quantitative trading models can provide meaningful alpha, particularly for stocks with higher retail participation. We discuss data preprocessing techniques, signal construction methodologies, and practical implementation considerations for integrating alternative data into systematic trading strategies.

Full Paper

Tip: Use your browser's zoom controls or the PDF controls to adjust the view. You can also scroll within the document or swipe on mobile devices.