This is a bit of an odd question but I was thinking about this for some time now and I would be very interested in best practice of doing this:
Let's assume I have a dataset with a time series variable X. Now I want to find ALL google search terms Y that have a (maybe spurious) correlation within the same time span with X. The problem is that you don't know which search terms were entered in google and therefore can't check for ALL correlations. So basically I want to do something similar to spurious correlations.
How would you start doing this?
I got this idea while using Google Trends
This was available as Google Correlate, but was recently deprecated.
You can read more about it in the whitepaper