Tech stack: n8n, X API, OpenAI, Supabase

Description:

Built an automated brand monitoring system that continuously tracks mentions across X (Twitter), enriches them with AI-powered sentiment analysis, and stores structured data for downstream reporting and response workflows.

The pipeline runs daily and handles the full lifecycle:

Raw mentions are pulled via the X API across multiple search queries, normalized into a consistent schema, and deduplicated against previously processed tweets to avoid redundant records. Each new mention is then passed through an OpenAI relevance check that classifies sentiment (positive/neutral/negative), identifies mention type, assigns a confidence score, and generates a plain-language reason — filtering out noise before anything hits the database.

Qualified mentions land in a Supabase table with full context: author profile data, engagement metrics (likes, retweets, replies, views), tweet URL, image metadata, and three timestamps tracking when the tweet was created, processed, and inserted.

Outcome:

A zero-touch monitoring system that surfaces relevant brand conversations daily, pre-scored and ready for prioritization — replacing manual search with a structured, queryable dataset that feeds directly into response and reporting workflows.