Google searches predict market moves
nature.com
nature.com
Your title doesn't reflect the article's title or its content, and is wrong in any case. No source can reliably "predict" the market's moves. Some sources can be shown to be correlated with the market after the fact, but that's very different -- among other things, it can't be used to guide advance actions like a real prediction would.
My advice -- avoid associating the words "predict" and "market" in anything resembling serious writing.
Of course, any real opportunity for a profitable trading strategy using this information would be quantified and reflected in prices, so I don't expect the results to persist over time, but I found the findings interesting nevertheless.
The efficient market hypothesis (EMH) says the market can't be gamed or predicted. No one knows whether the EMH is a legitimate theory or has an effect on the real market, but one thing is certain -- if there was a way to reliably predict the market, it would collapse. The market's very existence depends on the idea that all investors are equal and no one can (legally) game the system.
> This article may have been looking at correlations after the fact ...
That's one of the trickier aspects of investing -- the regularity with which one hears of a correlation that pops up after the fact, and the implication that it might have been exploitable in advance. But ex post facto correlations are just that, nothing more. They even have a name for it -- Monday morning quarterbacking.