ShraddhaRajcoomar13/UngizweBrandSOS

Anonymous brand intelligence platform that transforms customer and employee experiences into structured sentiment signals and topic-based pain point scores for South African brands. # BrandSOS Anonymous brand intelligence platform that transforms customer and employee experiences into structured sentiment signals and topic-based pain point scores. ## Overview BrandSOS captures anonymous "cries for help" from customers and employees, then converts them into actionable brand intelligence using AI-powered topic extraction and sentiment analysis. The goal is simple: help organizations identify emerging issues before they become crises. ## Current Features ### Frontend * Anonymous cry submission form * Brand selection and experience capture * Responsive web interface ### Backend * PHP form processing * MySQL data storage * Secure environment variable management * Python processing pipeline ### AI & Analytics * OpenAI-powered topic extraction * VADER sentiment analysis * Brand-topic aggregation * Topic-level pain point scoring ## Architecture Frontend → PHP → MySQL (`cries`) ↓ Python Intelligence Layer ↓ OpenAI Topic Classification ↓ VADER Sentiment Analysis ↓ `brand_topic_scores` ↓ Brand Insights ## Database Design ### cries Stores raw anonymous submissions. | Column | Description | | --------- | --------------- | | id | Primary key | | timestamp | Submission time | | brand | Brand name | | cry | Raw submission | ### brand_topic_scores Stores aggregated intelligence. | Column | Description | | -------------- | -------------------------- | | brand | Brand name | | topic | Extracted topic | | num_supporting | Number of supporting cries | | score | Topic pain score | ## Build Log ### Part 1 * Website prototype * Cry submission workflow * Database design * Brand intelligence concept ### Part 2 * MySQL integration * Python environment setup * OpenAI integration * Topic extraction pipeline * Sentiment analysis foundation * Brand-topic scoring architecture ## Roadmap * [ ] Real-time processing * [ ] Topic clust …