Business Intelligence - Predicting Threndse Present with Google T

Business Intelligence

Rationale behind the approach presented

In the April 2009 article “Predicting the Present with Google Trends” by Hyunyoung Choi and Hal Varian1, the authors’ view point is that the data reports provided by government agencies that economists, investors, and journalists so avidly follow are available after some delay, for example, data for a particular month may not be released until the middle of the next month. The authors think that using Google Trends may be beneficial. Google Trends provides daily and weekly reports on the volume of queries related to various industries. These query data may be correlated with the current economic activity to improve the prediction of subsequent data releases. The authors do not claim that Google Trends can help predict the future, but rather help predict the present in that knowing the query data in the middle of a month may help predict the sales in that month when it is released in the next month.

Business Applications

One of the main areas in business where the above data on Google Trends can be useful could be in marketing and sales. Knowing what the main queries being searched for can give an indication to a company about how to market its goods and products in a particular geographic area. The authors also provide several examples from various industries such as the automotive sector, retail sales, home sales, and travel. In each of these examples, the authors created models using econometric techniques called the seasonal autoregressive model. They used data that was provided by various government agencies. Additionally, the authors also included data from Google Trends in their models. Next, the authors performed several one-month ahead predictions using both the kinds of models and computed their mean absolute error. Each forecast used data available up to the point in time the forecast was made, which was one week into the month in question. The authors found that sometimes their models’...