PIA: Automated Price Intelligence Dashboard for Turkish Fashion Retail

Turkish fashion retailers frequently adjust prices across thousands of SKUs — making it difficult to tell whether a displayed discount reflects a genuine price drop or simply an inflated original price.

To investigate this, I built PIA (Price Intelligence App) — an end-to-end automated pipeline that scrapes publicly listed prices and product data across 7 categories daily using Python and Playwright, stores every price snapshot in a Supabase PostgreSQL database, and surfaces the data through a deployed Next.js dashboard. Discounts are calculated against each product's 30-day price high rather than the store-reported original price, exposing artificial markups that would otherwise go unnoticed.

The dashboard tracks 1,800+ products across three stores, with real-time filtering by store and category, a price history chart per product, and analytics including discount coverage, price segment distribution, and category depth comparisons. The scraper runs once a week via GitHub Actions and the dashboard is hosted on Vercel.