Spiral Graph: Cluster Buster

Pairing Citizen Scientists with A.I. to Identify Individual Arms in Spiral Galaxies

Alp Tezbasaran

alptezbasaran@ncsu.edu

Dr. Patrick Treuthardt

patrick.treuthardt@naturalsciences.org

Ian Hewitt

ihewitt@coastal.edu

2025-09-27

Our new Citizen Science Project

How did we end up here?

• Looking for galactic intermediate mass black holes
• Pitch angle is the key parameter, correlates with black hole mass

  

• So-called global pitch angle value might be misleading for galaxies with multiple arms

• Arm clustering is a way to identify individual arms

  

• We need to do this for thousands of galaxies!
• To scale up, we need a way to identify individual arms automatically
• AI can help us with this!

Arm Clustering

Training on good and bad clustering!

Predictions for new tracings!

So far, we have manually set hyperparameters

Meta-model for our hyperparameters!

Final Remarks

• Spiral Graph: Cluster Buster lives in Zooniverse
• Currently at demo phase
  • Instructions are complete
  • Tested with multiple volunteers
• Supported by NASA Citizen Science Seed Funding Program

Questions?

https://go.ncsu.edu/cluster-buster

THANK YOU

alptezbasaran@ncsu.edu

Pitch Angle correlates with BH mass

• Measure degree of winding (pitch angle) via P2DFFT
  • treuthardt.github.io/P2DFFT
• Pitch angle correlates with:
  • Maximum rotation velocity
  • Bulge stellar mass
  • Dark matter content
  • Supermassive black hole (SMBH) mass

• Origins of SMBHs are still unknown
  • Stellar BH rapidly consuming material?
  • Merging clusters of stellar BHs?
  • Collapse of large gas clouds?
    

Spiral Graph

Untraceable

Traceable

Interactive Tracing

• Identify the center galaxy
  • (a)Smooth, merging, or untraceable?
  • (b)Structured and not merging?
• If (b), trace the structure

Original Galaxy

Traced Arms