FluxAtlas

Inspiration

The world is full of resources, but not access. Critical materials (like metals, wood, minerals, agriculture inputs) are unevenly distributed across countries. Some nations have abundant resources but lack buyers or logistics. Others have high demand but no transparent way to acquire what they need. This creates inefficiency, price volatility, and lost economic opportunity. We were inspired to create a system that connects supply and demand fairly, regardless of geographic or economic advantage.

What it does

FluxAtlas is a global resource marketplace simulation platform. Countries (or agents) enter auctions to buy/sell resources. We cluster them based on PPP (purchasing power parity) so that bids remain fair, then use our pricing logic (Laplace bidding + Vickrey winner selection) to determine who gets the resource without letting the richest agents dominate.

In simple terms:
FluxAtlas matches underutilized resources to where they create the most value.

How we built it

• Modeled global resources using Postgres + SQLAlchemy
• Created PPP-based clustering to group countries by relative economic capacity
• Developed a bidding engine using:
  • Laplace-based bidding logic (to calculate willingness-to-pay fairly)
  • Vickrey Auction (winner pays second-highest price)
• Built a simulation interface that visualizes auctions, prices, and allocation outcomes

Challenges we ran into

• Building the Atlas engine from Scratch: simulating real marketplace behavior (dynamic bids, supply/demand shifts, pricing logic)
• Ensuring economically weaker bidders still have leverage in auctions

Accomplishments that we're proud of

• Successfully simulated cross-country resource trading with auction fairness
• Implemented Purchasing Power Parity clustering + Laplace bidding + Vickrey auction within a backend system
• Demonstrated how better allocation increases economic value for both sides

What we learned

• Efficient markets are not always fair, fairness must be designed intentionally
• Data modeling + economics + behavioral incentives = complex but rewarding
• Auction theory has real-world impact when applied correctly

What's next for FluxAtlas

• Integrate real global datasets (purchasing power parity, resource exports, trade flow)
• Explore partnerships with logistics and freight APIs to simulate delivery routing
• Introduce NUS (National Utility Score), a scoring model that quantifies each deal’s welfare impact, fairness, resilience, and sustainability benefits to ensure trades create real value, not just profit.

Built With

• fastapi
• postgresql
• python
• sqlalchemy
• tailwind