Interplanetary Path Evaluator

Challenges and Solutions

Mineral Composition
An initial challenge we faced was mineral composition, as we have little way of knowing the state of planets millions of light years away.
To solve this, we used a regularized linear regression (since a standard linear regression gave unstable weights) to assign weights to different metrics contained within the Exoplanet dataset.
After performing a Leave-One-Out (LOO) cross-validation to ensure accuracy, we applied the resulting weights to each of the 10 exoplanets.

Narrowing Down Exoplanets
Another challenge was deciding how many exoplanets to focus on. We knew we could not include every single exoplanet.
We chose 25 exoplanets to simplify the demonstration. In the future, we plan to implement a more concrete filter by establishing a Depot Readiness Score as a minimum requirement, rather than relying on an arbitrary cutoff.

Reflections

This was probably the most application-based project our group had ever done. Previously, we had learned coding and dataset analysis in high school classes, but actually applying those skills felt very different.

We are very proud of the way we took an idea and made it into a reality. At first we were a little disorganized, but we quickly realized the importance of clear role delegation and communication at every step of the process.

This idea is one we absolutely want to continue developing. We plan to implement the aspects of the 3D pathfinding program we left unfinished, and we’re excited about its potential for applications outside of space as well.

Overall, our team is hyped to see this through! 🚀