Personalized Video Streaming Platform

A full-stack video publishing and streaming platform featuring intelligent popularity prediction, seamless uploads, and a smooth user experience.

---

Features

Video Upload & Storage

• Secure video upload system using Node.js, Express, Multer
• Local storage management with automatic folder creation
• MongoDB integration for storing video metadata
• CORS-enabled API for frontend–backend communication

Machine Learning Video Popularity Prediction

• Data preprocessing using pandas
• Popularity classification using RandomForestClassifier
• Automatic labeling of videos as popular or not popular
• Accuracy reporting and prediction export

Personalized Content Visibility

• Recommends videos labeled as popular
• Improves user engagement through predictive analytics
• Helps creators understand performance based on past data

---

Machine Learning — Popularity Prediction

Tech: Python, pandas, scikit-learn

The ML pipeline:

1. Load and clean video data (Date.csv)
2. Remove missing fields
3. Automatically label a video as popular if it has more than 10,000 views
4. Train a RandomForestClassifier based on likes
5. Predict popularity for all videos
6. Output the most popular videos

Backend — Video Upload Service

Tech: Node.js, Express, Multer, MongoDB, Mongoose

The backend handles:

• Video upload
• Metadata storage
• Server file management
• Database connectivity
• API endpoints for the frontend