We replicated some of the results from the paper Predicting the Present with Google Trends. We focused on 3.1 Motor vehicles and parts and were able to replicate everything in that session using original data. We also explored some extensions of the session, such as replicate the results with data from different time periods, evaluate different models, and forecast. We found a model that that can forecast sales for a whole year by using only previous years' data. We tested the model, and the R^2 ranged from 0.88 to 0.97 for the years 2014-2019. We made our predictions for the year 2020. A flexdashboard that helped us evaluate and visualize different models was developed.
Link to the file: Replication and Extensions of the paper "Predicting the Present with Google Trends Session 3.1"