CMSACamp 2021

Instructors

• Assistant Director, Lead Instructor: Ron Yurko (ryurko@andrew.cmu.edu) is a PhD student in the Department of Statistics & Data Science at Carnegie Mellon University and a part-time Data Scientist at Zelus Analytics. He previously interned with the Pittsburgh Pirates and worked in finance modeling risk. He has publications in statistical genetics and sports analytics, and is interested in methodological research for model-based clustering. He is a co-organizer of the annual Carnegie Mellon Sports Analytics Conference and Reproducible Research Competition, and has developed multiple R packages to enable easy access of publicly available data, such as nflscrapR.
• Teaching Assistant: Beomjo Park (beomjop@andrew.cmu.edu) is a PhD student in the Department of Statistics & Data Science at Carnegie Mellon University. His research interests largely lie in the robust statistical inference including how to better accommodate the model misspecification. He has also collaborated with oceanographers on developing statistical framework for large-scale oceanographic data. He previously worked on wide-ranging applications and extensions in Bayesian semiparametric regression and Variational inference while pursuing an M.Sc. in Statistics at Korea University, where he also earned a B.Sc. in Industrial engineering and Statistics.
• Teaching Assistant: Nicholas Kissel (nkissel@andrew.cmu.edu) is a second year statistics PhD student at CMU. He received an MS in statistics and BS in math & statistics from Pitt in 2019. He is am primarily interested in developing methods for generating model confidence sets. More broadly, he is interested in creating inferential procedures for machine learning modeling methods, as well as developing accessible statistical tools that are applicable to natural and social science research.
• Teaching Assistant: Meg Ellingwood (mellingw@andrew.cmu.edu) will be joining the department as a first-year PhD student this fall. She graduated from Kenyon College in December 2020 and spent a term teaching statistics at Peddie School in New Jersey in spring of 2021. Meg participated in CMSACamp last summer and is very excited to be involved again this year. She is interested in applications of statistics and data science in sports and neuroscience.

Program description

The program will concentrate on statistics and data science methodology with applications in sports. Details regarding program topics are presented below, however this calendar is subject to change. All lecture slides will be available at the Lectures tab, and all lab / demo materials will be available at the Labs tab.

Week 1: June 1 - 4

• Introduction to the program, department, CMSAC
• Exploratory data analysis, tidyverse
• Data visualization, grammar of graphics, ggplot2
• Using Git and GitHub

Week 2: June 7 - 11

• More data visualization, heatmaps, density estimation
• K-means and hierarchical clustering
• Gaussian mixture models

Week 3: June 14 - 18

• Linear regression
• Generalized linear models (GLMs)
• Model assessment, bias-variance tradeoff

Week 4: June 21 - 25

• Variable selection
• Penalized regression, Ridge, LASSO, elastic net
• Dimension reduction, principal component analysis (PCA)

Week 5: June 28 - July 2

• Smoothing splines
• Additive models

Week 6: July 5 - 9

• Decision trees
• Random Forests
• Boosting
• Neural networks

Week 7: July 12 - 16

• Special topics and projects

Week 8: July 19 - 23

• Special topics and projects

Week 7: July 26 - 30

• Project presentations