Learning to predict future Bitcoin value with RNNs

In this recipe, we will learn how to predict future Bitcoin value with an RNN. The key idea is that the temporal sequence of values observed in the past is a good predictor of future values. For this recipe, we will use the code available at https://github.com/guillaume-chevalier/seq2seq-signal-prediction under the MIT license. The Bitcoin value for a given temporal interval is downloaded via an API from https://www.coindesk.com/api/ . Here is a piece of the API documentation:

We offer historical data from our Bitcoin Price Index through the following endpoint:
https://api.coindesk.com/v1/bpi/historical/close.json
By default, this will return the previous 31 days' worth of data. This endpoint accepts the following optional parameters:
?index=[USD/CNY]The index to return data for. Defaults to USD.
?currency=<VALUE...