How ymumberson/tailwind-portfolio works
Project Type and Purpose
• This is a full-stack (front-end-centric) static/dynamic site built with Next.js, styled with Tailwind CSS, and deployed on Vercel.
• It serves as a personal portfolio and sandbox for small web projects (e.g., digit classifier, tic-tac-toe, weather app).
File Structure
• src/app – App Router layout, global CSS, shared React components, and page definitions.
– components: reusable UI pieces (Card, Navbar, Project)
– sections: page sections (HeroSection, CommitGraph, Experience, Overview)
– projects: individual interactive demos under their own directories
• public – static assets (images, TensorFlow.js model files for digit classification)
• Configuration – next.config.ts, tailwind.config.ts, postcss.config.mjs, tsconfig.json
• Deployment – no custom backend; Next.js handles SSR/SSG and client-side hydration; Vercel hosts.
README
• Confirms Next.js + Vercel, plus small interactive demos.
• No external internal database; any data (e.g., weather) is fetched at runtime from public APIs.
Main Components
a. Next.js Server/Build Pipeline
– SSR/SSG process, routing via src/app
– Build step (PostCSS, Tailwind)
b. Client (Browser)
– Renders React components, executes interactive logic (TensorFlow.js, tic-tac-toe state, weather API fetch)
c. Public CDN (Vercel Hosting)
– Serves built pages & static assets (SVGs, model.json, .bin files)
d. External Services
– Weather API (if used)
e. Static Asset Store
– Public folder for static ML models and images
Relationships and Interactions
• Build → Vercel CI/CD → Static & SSR pages → CDN
• Client HTTP GETs page → Next.js SSR/SSG → Browser
• Browser loads TensorFlow.js model from /public → runs inference
• Browser fetches weather data from external API
Architectural Patterns & Design Principles
• App Router (file-based routing)
• Component-based UI
• Static-first deployment with selective dynamic hydration
• Separation of concerns: “components” vs “sections” vs “projects”