Project 4 - Stock Predictor API

Overview

This project is a stock price prediction application built using LSTM (Long Short-Term Memory) neural networks to predict stock prices based on historical data. It uses Yahoo Finance API to fetch data and displays 7-day forecasts with intuitive graph visualization.

Project Structure


stock-predictor/
├── app/                # Flask app
│   ├── templates/
│   ├── static/
│   ├── main.py         # Flask routes
│   └── model.py        # LSTM logic
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
└── README.md

Docker Setup

The application is containerized using Docker to ensure consistent development and deployment. After cloning the repository, use the following commands to build and run the container:


            docker-compose build --no-cache
            docker-compose up

Once running, the app is accessible via http://127.0.0.1:5050. Below is an example of the terminal output after successfully running the application:

Flask API Structure

The core backend logic is built using Flask. It defines two main routes:

GET / - Renders the homepage with a form to enter a stock ticker
POST /predict - Receives the ticker symbol, triggers LSTM prediction, and returns results in JSON

The prediction results include:

predicted_prices: List of predicted closing prices for the next 7 days
volatility: Difference between max and min predicted prices
graph_url: URL to generated graph image
current_price: Latest real-time stock price

▲ main.py showing Flask route logic for receiving and returning predictions.

LSTM Prediction Model

The LSTM model is implemented in model.py using TensorFlow. It processes time-series stock price data and predicts the next 7 days.

Data is normalized and reshaped into LSTM-compatible sequences
The trained model outputs future price predictions
The results are saved and plotted using Matplotlib

▲ Key portion of LSTM implementation and prediction logic in model.py.

Key Features

7-Day Price Prediction: LSTM model forecasts upcoming prices
Graph Visualization: Visual output using Matplotlib
Volatility Measurement: Calculates and displays price fluctuation
Current Price Display: Shows most recent market price
Dockerized: Entire app runs inside a container

How It Works

The application fetches stock data using yfinance and trains a TensorFlow LSTM model. After prediction, it generates a graph and returns results via a Flask API.

GET /: Homepage with input for ticker
POST /predict: Returns JSON with prices, graph URL, current price, and volatility