PUNCH
Scientific Visualization Process
An independent workflow for reading, processing, enhancing, and animating publicly available data from NASA’s Polarimeter to Unify the Corona and Heliosphere mission.
Overview
PUNCH studies the Sun’s outer corona, solar wind, and the connection between the corona and heliosphere. Its public files are scientific datasets rather than finished conventional images.
I began with no previous experience working directly with PUNCH mission data. The project became an experiment in AI-assisted technical learning: inspect unfamiliar FITS structures, locate the science image planes, build a Python environment, correct failed assumptions, and create a repeatable visual sequence.
The goal was not to reproduce an official mission product. It was to develop an independent, transparent processing path and clearly separate scientific handling from interpretive presentation.
How It Started
The project began as an attempt to transform publicly available scientific data into something visually understandable without hiding the technical steps required to get there.
The process required learning how FITS containers were structured, identifying compressed image extensions, and determining which processing choices revealed structure without implying unsupported measurements.
Can I enter an unfamiliar scientific-data environment, build a repeatable processing workflow, and communicate the result without overstating what the visualization proves?
Architecture
Acquire + Inspect
- Download FITS products
- Preserve metadata
- Inspect HDUs
- Locate image planes
Extract + Normalize
- Convert to arrays
- Handle invalid values
- Percentile range
- Consistent scaling
Enhance
- Asinh stretch
- Median filtering
- High-pass detail
- Compare treatments
Output
- Processed PNGs
- Frame sequencing
- FFmpeg / ImageIO
- Interpretive presentation
Development Timeline
Project start
Public mission data was selected and the Python environment was assembled.
FITS inspection
The science image was located inside a compressed image extension with multiple large image planes.
Processing pipeline
Normalization, asinh stretching, filtering, frame consistency, and export were developed and corrected.
Animation and comparison
Processed frames were assembled into time-based output and compared with mission-defined visualization guidance.
Package and publish
Scripts, parameters, example files, attribution, and scientific/interpretive comparisons need final release packaging.
Results and Current Boundaries
Complete data-to-image path
The project produced a repeatable route from mission FITS files to processed stills and animation.
Technical learning demonstrated
An unfamiliar data structure and scientific Python environment were successfully inspected, repaired, and used.
Claim boundaries clarified
The project now distinguishes official products, reproducible reference processing, and independent interpretive visualization.
This is an independent visualization, not an official NASA, Southwest Research Institute, or PUNCH mission product. Additional visual processing should not be mistaken for new scientific measurement.
Project Materials
Processing
Python scripts for FITS inspection, normalization, filtering, and export.
Examples
Original views, stretched frames, high-pass comparisons, and animation output.
Documentation
Parameters, workflow notes, attribution, limitations, and reproducibility guidance.
Add your final GitHub, ZIP, diagram, video, or document links here only after each public package has been reviewed and sanitized.