For this task, you'll need to regress the change in the log oil price on the change in log copper price, the change in log dollar, and the change in the 10Y Treasury rate. We estimated the regression with no intercept using data from 6/1/2011 through 6/1/2014. Note that, while Hamilton used log changes for most of his variables, for the change in ten-year Treasury rates, he used straight differences, so that's what we did as well. You'll need to do an out-of-sample prediction to get estimated values (based on the regression) of the daily change in oil due to demand from 6/2/2014 onward