07PartialTime series

Bitcoin 30 days ahead with AutoTS

I downloaded 701 days of Bitcoin prices and started an AutoTS search for a 30-day forecast. I stopped the search after it had tried 137 models, so the forecast cell never ran.

Original tutorial by Aman Kharwal: Cryptocurrency Price Prediction for the next 30 days (opens in a new tab)

What it printed in 2022

Numbers read from the saved notebook outputs. Cell numbers count every cell in the notebook, Markdown included, starting at 1.
Output	Value	Cell #
Daily BTC rows downloaded	701	6
AutoTS models tried before the run stopped	137	11

Cell # counts every cell from the top of the notebook, Markdown included, starting at 1.

Where it stopped

The AutoTS fit was interrupted during its first generation and the prediction cell was never executed.

The data

Daily BTC-USD prices from Yahoo Finance through yfinance.

What I'd do differently

Give AutoTS a time budget (a short model list and fewer generations) before starting.
Backtest on several past windows and compare with a flat 'no change' forecast.
Show a prediction interval. A 30-day point forecast for Bitcoin hides how wide the plausible range is.