The Proper Use of Google Trends in Forecasting Models

TL;DR

This paper examines the variability in Google Trends data and its implications for forecasting accuracy, highlighting potential pitfalls and proposing methods to improve reliability in predictive models.

Contribution

It reveals the issue of sample variability in Google Trends data and offers strategies to mitigate its impact on forecasting models.

Findings

Google Trends data samples vary significantly over time

Sample variability can lead to arbitrary forecasting conclusions

Proposed methods improve the robustness of Google Trends-based forecasts

Abstract

It is widely known that Google Trends have become one of the most popular free tools used by forecasters both in academics and in the private and public sectors. There are many papers, from several different fields, concluding that Google Trends improve forecasts' accuracy. However, what seems to be widely unknown, is that each sample of Google search data is different from the other, even if you set the same search term, data and location. This means that it is possible to find arbitrary conclusions merely by chance. This paper aims to show why and when it can become a problem and how to overcome this obstacle.