StarCrossed

Inspiration

Space exploration is becoming more ambitious every year, with organizations like NASA, ESA, and private companies planning missions to the Moon, Mars, asteroids, and beyond. However, choosing the right destination for a mission is incredibly complex. Engineers and scientists must balance factors like radiation, gravity, water availability, temperature, travel distance, communication delays, scientific value, and cost.

We wondered: What if choosing a space destination felt as intuitive as using a dating app?

That idea became StarCrossed, a playful "space Tinder" where users create a mission profile and swipe through celestial bodies until they find their perfect match. Behind the fun interface is a decision-support system that compares real planetary data against mission priorities to recommend destinations for exploration, colonization, mining, or scientific research.

What it does

StarCrossed helps users discover the most suitable destination for their space mission.

Users create a personalized mission profile by selecting information such as:

Mission objective Budget Crew size Mission duration Scientific priorities Risk tolerance Resource preferences

The app then compares these requirements with real scientific data from planets, moons, dwarf planets, asteroids, and exoplanets.

Instead of simply displaying information, StarCrossed calculates a compatibility score and explains why each destination is a good match using factors like:

Gravity Radiation levels Water availability Temperature Distance Communication delay Scientific potential Human colony suitability (which was calculated through many different factors and publiclly available data) Resource availability

Finally, users receive a mission compatibility report summarizing the strengths and tradeoffs of their chosen destination.

How we built it

We built StarCrossed using:

Python Streamlit for the interactive web application Custom CSS for the space-themed dating app interface JSON datasets containing real planetary and astronomical information Session State to carry mission information across multiple pages

The application is organized into multiple pages:

Home Mission Profile Swipe Destinations Your Match Mission Report

Our recommendation engine compares mission preferences with planetary characteristics and generates a compatibility score based on multiple scientific factors instead of relying on a single variable.

Challenges we ran into

One of our biggest challenges was designing a scoring system that balanced multiple competing priorities. Different missions require different environments, so a destination that's excellent for robotic exploration may not be ideal for a human colony.

We also had to redesign parts of the application after expanding our planetary database from just a few planets to dozens of planets, moons, dwarf planets, asteroids, and exoplanets. Updating the data structure meant modifying several pages so they all interpreted the new information correctly.

Finally, building a polished multi-page Streamlit application within only a few hours required careful organization, debugging, and rapid iteration.

Accomplishments that we're proud of

Creating a unique dating app experience centered around space exploration. Building a complete multi-page application in only four hours. Using real astronomical data instead of fictional planet descriptions whenever possible. Designing an interactive recommendation engine rather than a static encyclopedia. Creating a fun interface that makes complex aerospace concepts accessible to everyone.

What we learned

This project taught us that recommendation systems can be applied far beyond shopping or social media. Even space mission planning involves comparing many variables simultaneously and making tradeoffs between safety, science, cost, and feasibility.

We also gained experience working with:

Multi-page Streamlit applications JSON data management State management User interface design Decision-support algorithms Scientific data visualization

Perhaps most importantly, we learned how to quickly prototype an idea while continuously improving it through testing and debugging.

What's next for StarCrossed

We're excited about several future improvements:

Integrate a large language model to generate personalized mission briefings and scientific explanations. Connect directly to public NASA and ESA APIs for continuously updated planetary data. Replace the simple scoring system with a weighted machine learning recommendation model. Add interactive 3D planetary visualizations and orbital maps. Include radar charts comparing multiple destinations side by side. Support mission planning for real NASA mission concepts and future Artemis and Mars exploration scenarios. Allow users to build collaborative missions with teammates and compare different mission strategies.

StarCrossed began as a fun "space Tinder" concept, but it demonstrates how AI-assisted decision support can make complex scientific planning more intuitive, engaging, and accessible to students, educators, and future space explorers alike.