In order to build Regression or Deep Learning models for predicting the market, we need a bunch of historical data.

Prices and technical indicators are easily accessible, but getting news from the web or using the frequency of keywords searched at google as input data is much more complex.

Are there any available datasets with information where we can use to train our models? Kaggle had a competition using a similar dataset but now the competition is closed and can not be accessed.

Alternatively, is there a clever way to draw such data?

Thank you!

  • 1
    $\begingroup$ might be a duplicate question of getting sentiment data and leads contained there: quant.stackexchange.com/questions/24591/… $\endgroup$
    – develarist
    Commented Nov 29, 2019 at 15:19
  • $\begingroup$ As there are papers on this topic, my suggestion would be to read those and see what sources they used. From memory, Refinitiv (formerly Reuters) is one but it will likely be expensive. $\endgroup$
    – user42108
    Commented Mar 18, 2021 at 15:38


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.