Kan Nishida, a CEO of explanatory.io, carried out a research on whether there is connection between price of Bitcoin and activity on Google trends. He published his results on Medium.
Firstly, he studied correlation between price and google trends. The correlation value was 0.95 what allowed the author to say there is high connection between two variables. However, correlation does not identify causality; Nishida continued by running regressions.
He then built a linear regression model to predict BTC price by Google trends. The R-square of the model was 0.89 what meant high explanatory power. The coefficient estimated was shown to be statistically significant and positive. Hence, the more Bitcoin-related activity is Google Trends is seen the higher should be the price of BTC.
In conclusion, author states again that looking only at correlation is useless as it does not provide researcher with a direction of relation between variables. He also marks that other factors should be taken into consideration while developing your own predicting models.