S&P 500 Stock Movement Predictor

Project Overview

Developed a dynamic stock prediction model that analyzes S&P 500 stocks using Auto ARIMA (Autoregressive Integrated Moving Average) modeling. The system identifies top-performing stocks based on historical returns and generates future price predictions.

Python

Pandas

pmdarima

Yahoo Finance API

Matplotlib

NumPy

Key Features

Dynamic Stock Selection

Automatically scrapes and analyzes all S&P 500 stocks, ranking them by historical performance to identify top candidates for prediction.

Flexible Forecasting

User-configurable number of stocks to analyze and prediction timeframe, allowing for customized portfolio analysis.

Advanced Analysis

Implements Auto ARIMA modeling with automatic differencing and parameter selection for optimal predictions.

Workflow

Scrapes current S&P 500 stock list from Wikipedia
Downloads historical price data using Yahoo Finance API
Calculates and ranks stocks based on historical returns
Selects top N performing stocks (user-defined)
Performs autocorrelation analysis
Generates return distribution histograms
Applies Auto ARIMA modeling for each selected stock
Produces visual forecasts with confidence intervals

Visual Outputs

For each analyzed stock, the model generates:

Autocorrelation plots showing temporal relationships in price movements
Return distribution histograms
Price prediction charts showing:
- Historical trained data (blue)
- Actual prices (red)
- Model forecasts (green)
- Future predictions (yellow)

Model Details

Uses pmdarima's auto_arima for automatic model selection
Implements automatic differencing using KPSS and ADF tests
Training/test split: 80/20
Evaluates predictions using Mean Squared Error and SMAPE
Updates model parameters with each new observation

Try It Yourself

Run Stock Predictor in Google Colab

What You Can Do:

✓ Choose how many top S&P 500 stocks to analyze
✓ Set your prediction timeframe
✓ See performance rankings and predictions
✓ View detailed visualizations and analysis

Just click the button above and run all cells in the notebook to get started!