← back

PremeBase | StockX Analytics Platform

2019Reseller Era

StockX analytics platform. Web scraping pipeline, historical price tracking by model/size, and trend analysis models.

The Problem

StockX and GOAT have opaque pricing. Historical sale data is behind paywalls or requires manual searching. For sneaker resale, understanding price trends was essential for profitable flipping.

Technical Approach

Data collection and analytics platform:

- Web scraping of StockX sale listings

- Historical price tracking by model and size

- Trend analysis and price prediction models

- REST API for programmatic access

Interesting Challenges

StockX's anti-scraping measures required careful request pacing and header management. I learned about rate limiting, IP rotation, and session management.

The analytics were useful but the predictions were naive. Price prediction in resale markets is heavily influenced by hype and influencer behavior that statistical models can't capture.

What I'd Do Differently

The project was technically functional but didn't produce actionable predictions. The real value would have been in faster data collection or better data sources, not improved models. I pivoted to baseball modeling where the underlying phenomena are more predictable.

Tech Stack

PythonFastAPIPostgreSQL