PHYS 576 A: Selected Topics in Experimental Physics

Zoom Link https://washington.zoom.us/j/96898385387

Welcome to Modern Data Analysis Techniques

Team taught by Miguel Morales (Physics) and Bryna Hazelton (eScience), the goal of this class is to introduce current techniques and best practices in the statistically rigorous analysis of large data sets. The class is organized around four themes:  practical statistics, advanced data visualization, building collaborative analysis code, and advanced data analysis practices.

Grading

As a graduate elective, what you get out of the course largely depends on what you put into it. Further, this class is designed to scale depending on your interests and time. At one end, it is designed to provide motivated students with a firm grounding in advanced statistics and data analysis tools that can be used on a wide range of academic and professional problems. At the other end it is designed to serve as a low-pressure survey of modern analysis techniques. During the first week you will detail what your goals are, and your grade will be based on how well you achieve your goals. There will be no exams, with the homework and final project forming the basis of your grade. 

Syllabus

(Lecture title links to zoom, link to slide pdfs follows. Syllabus still under development, subject to change.)

Week 1

T:  Welcome; course overview; what does sigma mean? slides 

Th:  Analysis chains; Introduction to git & GitHub  slides (analysis chain; git)

Homework: Intro quiz

Week 2

T:Statistical building blocks (slides; matlab pt1 & pt2)

Th: Data visualization pt. 1; slides

Homework: Homework #1 

Week 3

T:  Data visualization pt. 2 (more examples);  poisson ± sigma difference; trials factors; parameter distributions; slides

Th: Workshopping your plots; systematics scavenger hunt; time-position variable backgrounds ; slides

Homework:  HW #2 

Week 4

T:jackknife & statistically valid plots; examples; HW discussion ; slides

Th: Stats review & common errors;  python hints; slides

Week 5

T:  Developing an analysis plan:  statistical worries; git issues; slides

Th:  Confidence intervals; slides

Homework:  HW #3 

Week 6

T:  Metadata, Provenance & Test Thickets; slides

Th:  Machine Learning (the blob pt 1); slides

Week 7

T:  Parameters, inherited analyses (the blob pt 2); slides

Th:  Blind & semi-blind analyses (Hertzog guest)

Week 8

T: Plots as a language

Th (5/20): Presentations:  Michael Pun, Debby Tran, Samantha Tetef

Week 9

T (5/25): Presentations:  Anna Wirth-Singh, Chris Thomas 

Th (5/27): Presentations:  Samantha Gilbert, Miguel Morales

Week 10

T:  Presentations:  Tharindu W. Fernando; Data rampages

Th:  Presentations:  Rodolfo Garcia, Robert Pecoraro, David Wang

Holding pen (early):  example of multi-dimensional probability; multi-parameter distributions, multi-dimensional spaces and triangle plots

Holding pen (late):  art of parameters, blind & semi-blind analyses, peer reviewed code.