Degree Path

Project Story

What Inspired Us

As Lehigh students, we’ve struggled with confusing degree audits, scattered planning tools, and the constant fear of overloading a semester. We’ve seen friends delay graduation because of a missed prerequisite or a badly timed schedule. That frustration inspired us to build Degree Path Simulator: a single, AI-powered web app that turns course planning into a clear, visual experience instead of a stressful guessing game.

How We Built It

We formed a focused 2-person team: one leaning into product vision and UX, the other into AI tooling and full-stack implementation. In under 24 hours, we:

• Sketched the core flows: Course Catalog → Cart → Plan Ahead → Progression → Advisor / Registration.
• Scaffolded a React + TypeScript frontend with a modern component library for fast UI iteration.
• Set up a backend with a relational schema for students, courses, requirements, and plans.
• Used AI tools to:
  • Generate boilerplate components and layout skeletons.
  • Help with type definitions, API hooks, and edge cases.
  • Draft validation logic and copy for the UI.
• Wired the “AI plan generator” to take user context (major, credits, preferences) and propose a semester-by-semester path that avoids overload and respects prerequisites.

We didn’t need heavy math, but we liked thinking about each plan as optimizing a kind of “load function” over semesters, roughly balancing something like
$$\text{difficulty_score}\text{semester} \approx \sum{i} \text{credits}_i \cdot \text{course_weight}_i.$$

What We Learned

Over the hackathon, we learned:

• Scope is everything. We had to choose a thin but powerful slice: make planning feel intelligent, even if not every edge case is solved.
• AI is a force multiplier, not a replacement. AI sped up boilerplate (components, hooks, copy), but we still had to make the product decisions, data model, and user experience coherent.
• Good UX beats raw features. A clean flow from catalog → cart → plan → advisor is more valuable than throwing in ten disconnected AI tricks.

Challenges We Faced

We hit several challenges along the way:

• Data modeling under time pressure – deciding how to represent courses, prerequisites, and progress in a way that still allowed “what-if” simulation.
• Aligning AI behavior with real constraints – keeping generated plans realistic (no impossible schedules, no broken prereq chains) within the time limit.
• UI complexity – coordinating multiple views (catalog, cart, plan grid, progression flow) without overwhelming the user.
• 24-hour constraint – continuously deciding what to cut and what absolutely had to ship for the demo.

Despite these constraints, we ended with a working prototype that unifies degree audit, plan ahead, course catalog, and registration into one visual, AI-assisted planner—exactly the kind of tool we wish we’d had when we started our degrees.

Built With

• auth
• chatgpt
• codex
• github
• javascript
• lovable
• npm
• postgresql
• react
• sql
• supabase
• tailwind
• typescript
• vite