By obtaining data from Google Trend data, which forecast method should be used to predict housing prices and sales?

Many scholars and companies around the globe are using Google Trend data to conduct researches and to forecast about current and future economic activity. Google Trend data has helped a lot in providing query data to predict current and future trend related to various industries in a timely manner. Before 2004, there was no source to provide daily or weekly data, as the data needed times to compile, revised and not all data was available in all geographic areas. Thus, there was usually a time lag for the release of current data trend. Google Company had realized this problem and as a result, they introduced Google Trends to the public in 2004. Since then, Individuals and companies can obtain daily and weekly reports on the volume of queries related to various industries from Google Trends. Many researches have proven the effectiveness of using Google Trend data to predict about the present trend. For instance, a recent study by three economists had proven the usefulness of Google Trend data in predicting daily price moves in the Dow Jones industrial average (as cited by Leinweber, 2013). In this essay, I would like to explain one of forecast methods that can be used to predict about changes in housing prices and sales based on a research by Lin Wu and Erik Brynjolfsson (Lynn Wu; Erik Brynjolfsson, 2013).

One of the most widely used and popular forecasting techniques is a linear regression analysis. The research about how Google searches foreshadow housing prices and sales also used a linear regression analysis to predict about the future trend of housing market. From my point of view, a linear regression analysis is a more appropriate forecasting technique to predict housing sales and prices rather than any other forecasting technique since there are no patterns such as cycles or trends in housing query data and line when they are plot and the deviations around the line are also normally distributed (As shown in Figure 1). The linear regression analysis uses an independent variable to predict a dependent variable. If there is a strong or at least moderate correlation between an independent variable and a dependent variable, we can use changes in an independent variable to explain/predict changes in dependent variable. In the research about using Google Trends to predict housing prices and sales, the authors had used search frequency as an independent variable and used housing price and sale volume as an independent variable. The research had found out that there is a high correlation between search frequency and housing sales while there is a moderate relationship between housing related searches online and housing price. The authors also claimed that a linear regression analysis of using search frequencies data to predict future US home sales can even beat the forecasting made by experts from the National Association of Realtors. According to their findings, we can use a linear regression analysis of using data from Google Trends to forecast housing prices and sales. However, we should note that there are more errors in predicting housing prices than predicting housing sales with the use of a linear regression analysis of using Google Trends to make prediction.

            To sum up, Google Trends provide social science researchers to use query data to predict/forecast about current and future economic activity in a timely and cost effectiveness manner. Nowadays, researchers are no longer relied solely on costly, time-consuming surveys and census data to make prediction about what is happening and what will happen in the marketplace. Lin Wu and Erik Brynjolfsson had proven the effectiveness of using a linear regression analysis of using Google Trend in predicting housing market trends. Notably from their research, Google Trend data can even predict about housing market trends more accurately than experts.

Figure 1: Quarterly Search Index for “Real Estate” normalized to total search volume ranging from 0 to 100.

References

Leinweber, D. (2013, 04 26). Forbes. Retrieved 10 30, 2014, from Big Data Gets Bigger: Now Google Trends Can Predict The Market: http://www.forbes.com/sites/davidleinweber/2013/04/26/big-data-gets-bigger-now-google-trends-can-predict-the-market/

Lynn Wu; Erik Brynjolfsson. (2013). The Future of Prediction: How Google Searches Foreshadow Housing Prices and Sales.