This project leverages Prophet (by Meta) and YFinance to analyze and forecast the price of Bitcoin (BTC-USD). By processing historical data from 2019 through late 2025, the model identifies cyclical trends and provides a 365-day predictive outlook.
- Analysis: Python, Pandas, YFinance
- Forecasting Engine: Meta Prophet (Additive Regression Model)
- Visualization: Matplotlib
- Dynamic Extraction: Utilized the
yfinanceAPI to pull real-time historical data. - Index Flattening: Handled the
MultiIndexstructure of modern financial data to isolate 'Close' prices. - Format Standardization: Converted timestamps to a timezone-naive format and renamed columns to
dsandyfor Prophet compatibility.
1. Bitcoin Price Prediction (Next 365 Days)
This plot shows the historical Bitcoin price (black dots) and the model's prediction (blue line). The light blue shaded area represents the Uncertainty Interval, accounting for potential market volatility.
Insight: The model predicts a year-end target exceeding $150,000 for late 2026.
2. Long-Term Growth Trend (Trend vs. DS)
This component removes the "noise" of daily price swings to show the underlying direction of the market.
Insight: The trend shows a massive growth trajectory starting from early 2023 through the 2026 forecast period.
3. Weekly Market Pulse (Day of Week)
Identifies which days of the week historically show the most strength or weakness.
Insight: Bitcoin shows a significant relative strength peak on Wednesdays.
4. Yearly Seasonality (Day of Year)
Shows how the price fluctuates based on the time of year (e.g., Q1 vs Q4).
Insight: The model detects recurring price dips in January and peaks during the mid-year cycle.
5. Intraday Volatility (Hour of Day)
Tracks the price movement throughout a 24-hour cycle.
Insight: Micro-volatility is tracked by the hour, showing specific windows of higher activity.