Realtime Reviews Sentiment
Realtime Reviews Sentiment is a full-stack data app that scrapes live user reviews from Letterboxd, analyzes their sentiment using NLP (VADER), and displays insights on a Streamlit dashboard. It includes a FastAPI backend, a background worker that continuously collects and classifies new reviews, and an interactive dashboard for visualizing sentiment trends across movies.
Context
I wanted an end-to-end data project that works on live input rather than a static dataset: scrape real reviews, classify them, and watch sentiment move in real time.
Stack
PythonFastAPIStreamlitSQLAlchemyVADER SentimentBeautifulSoupRenderDockerSupabaseLinks
Architecture
The app is split into three parts: a FastAPI service, a background worker that scrapes and classifies new reviews continuously, and a Streamlit dashboard. VADER handles sentiment because it is light enough to run on every incoming review, and Supabase stores the results so the worker and dashboard stay decoupled.