1. Final thoughts
You've reached the end of the course. Congratulations! I want to give you some final advice that's very important.
2. Toy examples
First off, everything we did here was a toy example. We were using very small datasets so it could run quickly on your browser. Real-life data is going to be much larger, so may take longer to run. To run big data calculations faster, look into parallel processing through Python 3's multiprocessing library, the Dask library, and Spark, among others. You may also need to use a cloud computing service, like Amazon Web Services, if your local computer can't handle the data.
3. Get more and better data
Secondly we used very limited data -- just the price and volume. You can get some free data similar to what we used from quandl.com/EOD. This data is extremely easy to get, so if you can actually get a statistical edge from it trading in the markets, that's amazing -- but also unlikely. Many academic experts and Nobel laureates have said, and even demonstrated, that it's difficult if not impossible to predict future stock prices from previous price data alone.
If you want to use some of the things you've learned here, you should incorporate more data. "Alternative data" is a buzzword for some of this -- it includes satellite imagery which can be used to estimate sales from things like cars on lots or shipping containers, sentiment analysis from social media, among others. You may also want to look into using analyst target price predictions, fundamentals data, and any other data you can get your hands on.
4. Be careful, and Godspeed!
Lastly be careful! The stock market can be extremely risky, so triple-check everything and make sure to test on an unseen test set before deploying any strategies. Many experts in the field suggest only investing in low-cost ETFs, and advise against stock-picking. Whichever path you choose, good luck, and Godspeed!