University Project / CS 3620
Yelp Scraper to Simplified View
A Python and Django final project that used Selenium and WebDriver to collect web data, shape it into a cleaner structure, and present a simplified view.
I needed a manageable final project for CS 3620 Server-Side Architecture, so I chose a web scraper rather than expanding the assignment into a much larger application. The goal was to collect Yelp data and redisplay it in a cleaner, more focused interface.
The project became an early lesson in separating extraction from presentation: HTML gathered through browser automation was not automatically useful application data. It had to be inspected, organized, stored, and then shaped for the interface that would consume it.
01 / Scope
Keep the final project focused on the server-side flow.
A deliberately contained idea left room to practice the parts of the assignment that mattered most: gathering information, structuring it, and serving it through a web application.
Dynamic source
Selenium and WebDriver provided a way to work with content rendered on a live website instead of relying on a static HTML file.
Structured records
The scraped results needed a consistent shape before the Django application could store and use them cleanly.
Simplified output
The final view prioritized a smaller set of useful information instead of reproducing the source page around it.
02 / Data Path
Treat raw HTML as an input, not the final model.
The useful work happened between collection and display, where page content became records that the application could reason about.
Collect
Browser automation navigated the source and extracted the information needed for the project.
Organize
The collected values were sorted into a more predictable data structure rather than passed straight into the UI.
Present
Django rendered the cleaned data as a custom page designed around readability rather than source-site fidelity.
03 / Later Relevance
The same data-shaping problem returned at a larger scale.
The coursework was small, but its core lesson became useful later while working for Aurawave Studio on a production-facing directory platform.
Messy inputs
Both projects started with HTML-derived data that needed inspection before it could support a reliable interface.
A stronger model
The later Aurawave work required turning scraped Python data into a better database structure for a Next.js and TypeScript product.
Transferable practice
The class project gave me an early mental model for separating collection, data cleanup, storage, and presentation.