Task 1: Data Preprocessing and Exploratory Analysis

This task focuses on extracting, cleaning, and analyzing historical financial data for Tesla (TSLA), Vanguard Total Bond Market ETF (BND), and S&P 500 ETF (SPY) from 2015 to 2026.

Deliverables

• Jupyter notebook containing full EDA and visualizations
• Summary of data quality issues and mitigation strategies
• Stationarity test results with interpretation
• Key insights into asset behavior, volatility, and risk

Key Findings

• Tesla exhibits high growth and extreme volatility
• SPY provides stable, market-level returns
• BND acts as a volatility dampener
• Returns are stationary; prices are not

Task 2: Time Series Forecasting Models

This task focuses on predicting Tesla's future stock prices using both classical statistical and deep learning approaches.

Models Implemented

• ARIMA/SARIMA with automated parameter selection
• LSTM neural network using historical price sequences

Evaluation Metrics

• MAE
• RMSE
• MAPE

Key Findings

• ARIMA provides interpretable baseline forecasts
• LSTM captures nonlinear patterns and often achieves lower error
• Chronological train-test split prevents data leakage

Task 3: Forecasting Future Market Trends

This task extends the trained forecasting models to generate 6–12 month future price predictions for Tesla stock.

Key Outputs

• Multi-step future forecasts using LSTM
• Forecast visualization with confidence intervals
• Trend analysis and market interpretation

Insights

• Tesla shows a continued upward trend with increasing uncertainty
• Confidence intervals widen over longer horizons, reducing forecast precision
• Short-term forecasts are more actionable than long-term point estimates

Task 4 – Portfolio Optimization

This task applies Modern Portfolio Theory (MPT) to construct an optimal portfolio using a forecast-based expected return for Tesla and historical returns for SPY and BND.

Outputs

• Efficient Frontier with optimal portfolios
• Covariance matrix heatmap
• Maximum Sharpe Ratio portfolio weights
• Expected return, volatility, and Sharpe ratio

Key Insight

Combining forecast-driven expectations with historical diversification assets improves risk-adjusted portfolio performance.

Task 5 – Strategy Backtesting

This task evaluates the optimized portfolio against a 60/40 SPY–BND benchmark using an out-of-sample backtesting window (2025–2026).

Key Outputs

• Cumulative returns comparison
• Total return, annualized return, Sharpe ratio, and max drawdown
• Strategy viability assessment

Key Insight

The model-driven portfolio demonstrates improved risk-adjusted performance, but results should be interpreted cautiously due to simplifying assumptions.