Data Science Blog
This blog is being constructed to showcase and discuss my completed and ongoing data science projects. At the moment I am working on my backlog of projects that I haven’t written about yet.
Predicting Galaxy Masses and Binary Black Hole Fraction
When astronomers observe galaxies through telescopes they have to predict the masses by observing the spectral properties and using Stellar Population Synthesis (SPS) models to determine their composition. In this project, I worked with a couple of students at UNC to improve upon these models by including binary star interactions. The model is a Bayesian model where we calculate quantities by Monte Carlo sampling. We were also interested in how often we would expect to find large binary black holes, capable of producing LIGO events, in the relatively numerous dwarf galaxies.
Project Site
RESOLVE Survey
Gravitational Waves Detected
Titanic: Machine Learning from Disaster
To get familiar with Kaggle competitions I worked on a project the initial tutorial project. The goal is the predict who onboard survived the accident. In the initial project we wanted to see how much the predictions would change when the input data was scaled properly as opposed to unscaled (violated the assumptions of the underlying SVM model). We saw an approximately five percent improvement in accuracy by preprocessing the data properly.
Project Site
Kaggle Site
Feature Scaling
Notes
General thoughts and information I have found useful and want to save. This is typically unrelated to a project.