Social Comments Scraper

A full-stack web app that pulls comments off social posts and turns the raw, inconsistent responses into a clean, queryable dataset. The front end is a React + TypeScript console for launching jobs and reviewing what came back.

What it does

• Drives jobs from the browser. The React/TypeScript UI lets me kick off a scrape, watch it run, and review the harvested comments without touching the command line.
• Survives flaky endpoints. Social endpoints paginate oddly, rate-limit, and change shape, so the scraper handles pagination, backs off when throttled, and retries failed requests instead of dropping the run.
• Normalizes the mess. Comment payloads arrive inconsistent and deeply nested; I map them onto a single consistent schema so every record looks the same downstream.

Why it exists

• Exports for real analysis. The point is structured output, not a screen of text — clean, schema'd data you can hand to downstream analysis or load elsewhere.
• At scale, unattended. Resilience and normalization are what let a job grind through a large post without constant babysitting.

Stack

React + TypeScript front end driving a resilient scraping pipeline (pagination, rate-limit backoff, retries) with schema normalization and structured export. Deployed live.