Movie Identification Capstone

Motivation

This senior capstone required us to ship something meaningful on top of Elasticsearch. My team built the exact tool we wanted when we can’t remember the movie title but remember quotes, plot twists, or a couple actors. The goal was to show how relevance scoring plus external APIs can recreate the “it’s on the tip of my tongue” moment.

Project Overview

Users type whatever details they remember—plot fragments, cast, settings—and the React client calls an internal API backed by Elasticsearch. Results come from a curated movie dataset and are enriched with TMDb/IMDb metadata so each card shows the poster, synopsis, cast list, and an embedded trailer pulled from YouTube search. The richer the description, the tighter the candidate list becomes.

Challenges

• Wired up the IMDb/TMDb APIs for posters and cast information while keeping request quotas in check.
• Used the YouTube Search API to automatically map matched movies to the most relevant trailer clip.
• Post-processed a large CSV export, trimming out anything pre-1950 plus Bollywood titles so the index focused on the intended audience and elastic scoring stayed accurate.

Tech Stack

React, Node/Express, REST APIs, Elasticsearch, TMDb/IMDb APIs, YouTube Search.