A machine learning tool that predicts the virality potential of a YouTube video before you post it.
Input a video title, tags, and optional thumbnail — get back a virality score from 0–100 plus actionable recommendations to improve reach.
- Analyses patterns from a dataset of viral YouTube videos
- Uses NLP to extract signals from titles and tags
- Uses computer vision (OpenCV) to analyse thumbnail features
- Outputs a score and specific suggestions
- Python · Flask · scikit-learn · Pandas · NumPy · OpenCV
- HTML/CSS frontend
Built at DATA-HACK 2026 (SUDATA × CommSTEM, University of Sydney)
Team Runtime Terror — Atharva Srivastava, Jasmine Chen