Predict Engaging/Viral Content on X (Twitter) Using Hybrid AI Models
A machine learning pipeline to predict engagement-optimized posts by combining BERT's NLP power and XGBoost's structured data efficiency.
This repository hosts an end-to-end ML system that analyzes historical X (Twitter) analytics data to:
- Predict which topics/content styles will trend in my niche
- Recommend high-impact posting strategies based on my high-performing content in the past
- Explain key drivers of high engagement rates (hashtags, timing, etc.)
- Hybrid Model: BERT (NLP) + XGBoost (structured data) ensemble
- Real-Time Integration: X API for fresh analytics ingestion
- Production-Ready: Feature store, model serving, and monitoring
- Explainable AI: SHAP values for post recommendations