Galaxy Spectra PCA Pipeline
Research with Prof. Wittman involving high-dimensional signal processing.
| Role: Student Researcher | Stack: Python, R, PCA |
[cite_start]Working under Professor Wittman, I built a reproducible pipeline to analyze high-dimensional astronomical data[cite: 29].
Methodology:
- Preprocessing: Processed 2,323 spectra, handling missing data and anomalous flux spikes. [cite_start]I modified the unsharp masking technique to strictly avoid division artifacts during normalization[cite: 32].
- Dimensionality Reduction: Built a PCA-based outlier filtering system. [cite_start]I produced interpretable principal components (PC1-PC3) to visualize clusters and performed regression analyses[cite: 33].
- [cite_start]Deliverables: Delivered a fully scripted, reproducible pipeline in Python and R, along with advisor-facing reports explaining the high-dimensional signals[cite: 33, 34].