Exploring a Quick Example: United States Box Office Data

This article examines the potential value of box office data collected from the Mojo website. The hypothesis is that box office performance might be counter-cyclical to the economy, potentially offering insights for forecasting economic indicators.

To begin the analysis, a search for "Mojo" in the database yielded several entries. The Mojo Gross monthly box office data was selected for further investigation.

The exploration process involved selecting the "explore" button while maintaining a 1-step forecast horizon. The aim was to assess whether treating box office data as an exogenous variable could improve 1-step univariate forecasts of time series within the database.

The search process, which took approximately one minute, analyzed 16,000 series with monthly frequency. The top result showed a correlation with the Consumer Price Index (CPI) for Dental Services in Sweden. While this specific correlation may be spurious due to the large number of series examined, a general trend emerged suggesting some predictive power of box office data over CPI.

Further analysis of the dominant topics among the top 500 entries revealed two main groups: CPI-related and labor force-related. The results for the labor force-related filter are shown below.

The analysis suggests that the box office data may have some value for researchers focusing on labor force dynamics and price indices. However, the improvement in predictive power is modest, with R-squared improvements of approximately 3-5% using a simple model across a vast number of data series. This limited improvement could potentially be attributed to chance, suggesting that the data's value may be limited.

Despite these limitations, the data has been shared on the platform. This decision was made to ensure that researchers who find box office data to be a valuable exogenous variable in their analyses can easily locate it in their search results and contact the data owner for access.