Brush - Binging on a Rush

Ever want to binge fast, but spend more time deciding what show? Brush is a streaming optimizer that creates the perfect binge in seconds, based on your mood and schedule - no more endless browsing!

Comment 2

BRUSH

Streaming platforms offer too many choices. We’ve all spent more time scrolling than watching. With busy schedules and endless content, we wanted to build something that helps users get straight to the good stuff—based on how much time they have to binge and what genres they're looking for.

Brush generates binge-worthy playlists tailored to your mood and time constraints. Whether you want a 2-hour escape or a week-long plan, Brush delivers curated TV episodes that fit your schedule. It offers:

  • Quick Binge: a single-session plan.
  • Full Binge: a daily viewing schedule for your long-term entertainment.

As for the specifics of our model, here were the technologies we used throughout our workflow:

  • Data: Used IMDb datasets (title.basics, title.episode, and title.ratings) and merged them to get show metadata, episode runtimes, ratings, and popularity.
  • Feature Engineering: Calculated runtime closeness, normalized popularity (to avoid bias toward viral shows), ratings, and episode count.
  • Recommendation Engine: Designed a weighted scoring system:
  • Quick Binge: 60% popularity, 30% rating, 10% time closeness
  • Full Binge: 40% time closeness, 30% popularity, 20% rating, 10% episode count
  • Frontend: HTML/CSS with Flask backend.
  • Image Search: Used Google Custom Search API to fetch show thumbnails.

Challenges we ran into:

  • Cleaning and merging IMDb’s massive TSV files was more complex than expected.
  • Balancing quality, popularity, and timing in our scoring model took careful tuning.
  • Getting image search results to match show titles accurately with the API.

What we learned:

  • How to process and merge large real-world datasets.
  • Trade-offs in building recommender systems.
  • The importance of UX in making data-driven tools intuitive and enjoyable.

Our future goals:

  • Add user accounts and preferences to personalize suggestions over time.
  • Integrate streaming availability (e.g., what's on Netflix or Hulu).
  • Expand to include movies.
  • Implement AI-based mood detection from text or emojis.
  • Integrate it as a Chrome Extension so users can quickly access their options.

Built With

Share this project:

Updates