← Back to Spring into AI Updates

Calorie Tracker — Burn Calories, Not Tokens

Evan Rhea's Week 1 submission demonstrates how to combine Codex and GitHub Copilot to build a functional data visualization app. Python, FastAPI, React — and a cross-platform journey from Windows to Mac.

📺 Watch Evan's Demo

See the Calorie Tracker in action as Evan walks through his AI-powered workflow and demonstrates the app:

The Builder

Evan Rhea, a student at Iowa State, entered the Launch Into AI competition with a clear goal: build something functional using AI tools, focusing on data visualization over polished aesthetics.

AI Tool Workflow

Evan combined two powerful AI coding assistants to get this built:

GitHub Copilot OpenAI Codex Python FastAPI React

Cross-Platform Journey

One of the standout aspects of this submission was the portability challenge. Evan initially built the app on Windows, then used Codex to transcribe the code to work on Mac for the video demonstration.

This is a real-world problem that AI tools are increasingly good at solving — taking working code from one environment and adapting it to another without manual rewrites.

The App: Calorie Tracker

The app is simple and functional:

No complex auth. No database setup. Just upload data, see charts, understand your patterns.

What Makes This Submission Work

Evan's entry hits the competition theme squarely:

The Demo Moment

In the video, Evan demonstrates uploading a CSV file with sample fitness data and watching the charts populate instantly. He even does some on-the-fly analysis: "We're bulking because we're eating more than we're burning — about a 300 calorie surplus."

The app handles both bar charts (for comparing values) and line charts (for seeing trends over time). Week 1 and Week 2 show lower activity, then Week 3 ramps up through Week 6. Simple data, clear visualization, immediate insight.

In Evan's Words

"I wanted to take the GitHub Copilot that I already use, and I wanted to combine that with GPT's codecs. And I wanted to see like how that sort of agent to agent interaction would work... I think the whole thing took me about 20 minutes, max."

"It could look like a preschooler built it, which I specialize in... But a big part of the competition about also just making things exist, so not making them perfect."

The Lesson

This submission proves that you don't need to be a senior developer to ship something useful. With the right AI tools and a clear scope, you can go from idea to working app in a weekend.

The competition rewards shipping. Evan shipped.

🏆 Think You Can Build Something Better?

The competition runs until March 5th. Ship early, ship often, and remember: the multiplier rewards action, not polish.