stockPredictionApp(1.2.0)

A full-stack, machine learning–powered web application that forecasts stock price movements by combining historical data, financial sentiment analysis, and news headlines. Built with a Python/FastAPI backend and a Next.js frontend, the app delivers real-time predictions and sentiment insights to help users make data-driven investment decisions.

https://stock-prediction-app-two.vercel.app/

---

Table of Contents

1. Getting Started
2. Project Overview
3. Backend Features
4. Frontend Features
5. Tech Stack
6. Future Improvements
7. Credits

---

Getting Started

1. Clone the repository

git clone https://github.com/your-org/stockPredictionApp.git
cd stockPredictionApp

2. Backend setup

# 2.1 Create & activate virtual environment
python3 -m venv .venv
source .venv/bin/activate    # macOS/Linux
.\.venv\Scripts\activate     # Windows

# 2.2 Install Python dependencies
pip install -r requirements.txt

# 2.3 Launch FastAPI server
uvicorn main:app --reload --host 0.0.0.0 --port 8000

3. Frontend setup

cd frontend

# 3.1 Install Node.js dependencies
npm install

# 3.2 Start development server
npm run dev

---

Project Overview

Our Stock Prediction App combines financial data, news sentiment, and cutting-edge machine learning models to forecast future stock prices. It features:

• A FastAPI backend for data ingestion, feature engineering, model inference, and sentiment analysis.
• A Next.js/Tailwind-CSS frontend with intelligent search, interactive charts, and real-time prediction display.

---

Backend Features

• Stock Price Prediction:

  • Models: Linear Regression, Decision Tree, Random Forest, Gradient Boosting, XGBoost, CatBoost, AdaBoost, SVR, Stacking Ensemble
  • Trained on historical price & technical indicators to predict next-day direction (and price).

• News & Headline Sentiment:

  • Fetches recent headlines via NewsAPI.
  • Trains a headline sentiment model (~76% accuracy).

• Reddit Sentiment:

  • Uses PRAW to pull Reddit posts by ticker/keyword.
  • Classifies sentiment with TF-IDF + XGBoost (~76% accuracy).

• Financial Statement Analysis:

  • Retrieves annual reports via yfinance.
  • Correlates YOY metrics for feature selection.
  • Current model accuracy ~52% (work in progress).

---

Frontend Features

• Search Bar with Auto-Suggestion:

  • Powered by Finnhub API for real-time ticker/company lookup.

• Dynamic Prediction Dashboard:

  • Recharts visualizations of predicted vs. actual price.
  • Sentiment tag cloud & summary panels.
  • Sectioned view for each model’s output + price forecast.

---

Tech Stack

• Languages & Frameworks:

  • Python 3.9+ · FastAPI · Uvicorn
  • JavaScript/TypeScript · Next.js · React

• Data & ML:

  • scikit-learn · XGBoost · CatBoost · SVR · pandas · NumPy
  • yfinance · NewsAPI · PRAW

• UI & Styling:

  • Tailwind CSS · Recharts

• Deployment:

  • Render.com (backend) · Vercel (frontend)

---

Future Improvements

• Custom Timeframes: User-selectable 1D, 1W, 1M, 1Y ranges.
• Real-Time Streaming: WebSocket updates for live data & sentiment.
• Community Trending: Show what stocks are trending in community by using NLP on reddit posts or other posts.

---

Credits

• AnkurZing — Financial News Sentiment (Kaggle)
• US Stocks Fundamentals (Kaggle)

---