Context

I like the idea of building passive income. There are many ways out there and one conservative way that has worked since ages is to invest in dividend paying assets. But you never know which one to pick with a good risk/reward ratio. I thought it would be great to have all the data available to have a data driven decision.

Content

The data is a snapshot of all tradable instruments on robinhood.com by using their API. When I do my research I like to also go through the data on dividend.com, so crawled their website to merge it with the symbols I already got from robinhood.com.
I documented and published the script how to create a new snapshot in this repository. Please refer to the notebook if you want to create your own snapshot.
I recommend to use the pickle file (instruments.p) with pd.read_pickle('../input/instruments.p') because the data is already prepared and in the right shape and order. If you are an R user, you can still read the data with the instruments.json file.

Acknowledgements

Thanks to robinhood.com and dividend.com for providing this data publicly.

Inspiration

Which stocks do have the best risk/reward ratio?

Data description

Content

Checkout the repository for a detailed description of each column

This dataset aims to facilitate NLP sentiment analysis by investigating the correlation between news headlines and cryptocurrency prices. Cryptocurrencies were chosen as they have fewer external variables than traditional stocks, allowing for more accurate analysis.

The dataset was collected by web scraping headlines from the first page of Google News for various cryptocurrencies by date for several years. While headlines for many cryptocurrencies are included, the best results were obtained by limiting models to data from 2021 to the present, when cryptocurrencies became more mainstream.

Pricing data was gathered using the unofficial Robinhood API robin-stocks, and the begins_at column contains the corresponding date for each market price and headline. Note that the headlines are from that date, so to make predictions, the articles column needs to be shifted.