StarFinder

Inspiration

80% of the world's population has never seen the Milky Way due to light pollution.
We wanted to create a tool that makes light pollution data accessible and actionable — helping people discover dark skies and raising awareness about this often-overlooked environmental issue.

What It Does

StarFinder is an interactive web application that visualizes global light pollution using real NASA satellite data and the World Atlas 2015 dataset.

Users can:

Search any location in the U.S.
View detailed pollution metrics with our blue-to-red heatmap
Use our intelligent Dark Sky Finder to discover the best stargazing spots nearby

Our spatial distribution algorithm returns up to five locations spread across different compass directions, ensuring diverse options for finding pristine night skies.

How We Built It

We built StarFinder with:

Next.js 15 and TypeScript for the full-stack framework
Mapbox GL JS for interactive mapping
Tailwind CSS for the dark-mode UI

The backend uses Python scripts with rasterio and NumPy to process massive 3GB GeoTIFF satellite files into optimized JSON data.
We implemented custom algorithms for spatial search, haversine distance calculations, and bearing-based distribution to ensure intelligent result placement.

Challenges We Ran Into

Our biggest challenge was finding and processing real scientific data.
We initially struggled to locate publicly accessible, high-quality light pollution datasets.

After discovering the World Atlas 2015 dataset, we faced the technical hurdle of processing 3GB GeoTIFF files with over 750 million pixels into a format our web app could handle efficiently.
Converting this data while preserving accuracy and maintaining reasonable file sizes required multiple iterations of our Python processing pipeline.

Accomplishments We're Proud Of

We’re incredibly proud of successfully processing and optimizing real scientific data from the World Atlas 2015 dataset — transforming 3GB of raw satellite imagery into a streamlined 8.2MB JSON file with 50,000 globally distributed data points.

Our custom spatial distribution algorithm intelligently finds dark-sky locations across eight compass sectors with minimum separation thresholds — creating a genuinely useful tool rather than just a data visualization.

What We Learned

This project taught us how to work effectively as a team on a complex full-stack application.
We learned to divide responsibilities between frontend development, data processing, and algorithm design while maintaining clear communication.

We also gained hands-on experience with geospatial data processing, coordinate systems, and the challenges of optimizing large scientific datasets for web applications.

What's Next for StarFinder

We plan to integrate real-time weather and cloud cover data to help users identify not just dark locations, but clear nights perfect for stargazing.

Future features include:

Temporal analysis to track light pollution changes over time
User-contributed observations to build a community dataset
Astronomy mode with dark-sky park recommendations and constellation overlays for educational purposes