Stock trade volume prediction with Yahoo Finance user browsing behavior

Web traffic represents a powerful mirror for various real-world phenomena. For example, it was shown that web search volumes have a positive correlation with stock trading volumes and with the sentiment of investors. Our hypothesis is that user browsing behavior on a domain-specific portal is a better predictor of user intent than web searches. We focus on the financial domain and we analyze the web browsing and trading data of more than 2600 stocks traded on NYSE, Nasdaq and SNP. The web browsing data consist of user page views related to stocks on Yahoo Finance, while the trading data include the trading volume of these stocks. We study the correlation and causality between web browsing and trading data while varying the time granularity (hourly, daily) and financial segmentation (individual tickers, industries, sectors). We find that web browsing on Yahoo Finance can anticipate stock trading volumes by two or three days, resulting in a higher predictive power than previous work that used web searches to predict trading volume. We also observe that grouping stocks into industries or sectors decreases the predictive power, whereas moving from hourly to daily time series granularity improves predictive power. We corroborate our findings with a theoretical intuition and extensive statistical and causality tests.