Back to catalogue
Project
Tesla Random Forest Predictor
A Streamlit research demo that estimates the next closing price with a Random Forest.
How it’s builtHow it’s built
Tesla Random Forest Predictor · Streamlit demo predicting Tesla's next-day close with a Random Forest
I trained a 100-tree Random Forest regressor on daily TSLA OHLCV data from 2015 to 2024 pulled with yfinance to predict the next day's close, and built a Streamlit app where users enter today's open, high, low, close and volume. The app recomputes MSE, RMSE, MAE and R² live and falls back to stored metrics when Yahoo Finance rate-limits.
- input features (Open, High, Low, Close, Volume)
- 5
- trees in the RandomForestRegressor
- 100
- R² on a random, in-sample split (RMSE 6.68, MAE 3.45)
- 0.9965
- training data range
- 2015-2024
Technology and what it does here
- scikit-learn RandomForestRegressor
- Maps today's OHLCV to tomorrow's closing price
- yfinance
- Downloads TSLA daily prices for training and live evaluation
- Streamlit
- Input form, prediction and model performance table in app.py
- joblib
- Saves and loads tesla_rf_model.pkl (18 MB)
- pandas / NumPy
- Target construction with Close.shift(-1) and metric tables
Implementation details
- The app caches the model with st.cache_resource and the price download for an hour with st.cache_data(ttl=3600), and flattens yfinance MultiIndex columns.
- When Yahoo Finance is unavailable the app shows stored benchmark metrics instead of failing.
Technologies
- Random Forest
- Time Series
- Streamlit
