Estimating Hubble's constant (H₀) via photometric redshift through a CNN & a visualizer through ManimGL in Python.
This project was submitted as a final project for Physics 20B - Cosmology: Humanity's Place in the Universe at the University of California, Irvine.
HubbleVision.mp4
git clone https://github.com/sidsun1/HubbleVision.git
cd HubbleVision
Ensure that Python 3.12 is used for the virtual environment.
python -m venv venv
# Windows
venv\Scripts\activate
# Mac/Linux
source venv/bin/activate
pip install -r requirements.txt
Simply click play on the module, or run it with
python -m download_sdss_images.py
Click run all in the Jupyter notebook interface in the file redshift_cnn.ipynb.
Play the video files in the videos subdirectory, or run the commands:
python -m manimlib hubble_scene.py HubbleIntro -w --uhd
python -m manimlib hubble_scene.py HubbleLawComparison -w --uhd
python -m manimlib hubble_scene.py RedShiftScene -w --uhd
- https://skyserver.sdss.org/dr18/SkyServerWS/SearchTools/SqlSearch (Sloan Digital Sky Survey - Data Release 18) [Used for image downloads through SQL queries]
- https://www.tensorflow.org/guide/keras/sequential_model [Tensorflow CNN development docs to calculate Hubble's Constant]
- https://3b1b.github.io/manim/ [Docs for development of python animations on Hubble's Constant + RedShift / The Doppler Effect]
- https://cosmo.nyu.edu/mb144/manyd.html [Used to determine a standardized value for M_r]
- https://academic.oup.com/mnras/article/392/3/1060/1062734 [Also used to find a standard value for magnitude range]
- https://ar5iv.labs.arxiv.org/html/1811.04569 [Used to compare data based on parameters like ra, dec, and magnitudes across different bands]