Computational Statistics and Data Science Methods

Catalog description

This course will introduce traditional multivariate statistical methods and multi-model inference, as well as iterative computational algorithms (i.e. Bayesian methods and machine learning) appropriate for graduate students conducting or planning to conduct a graduate research project. The course will focus on the proper application of methods to a sample data sets using statistical programming software and graphics and will forego the more in-depth analytical mathematical exposition that you might see in a math course, so that we can cover a larger variety of methods and spend more time implementing them in code. Practical examples will often derive from the fields of biology, environmental science, or medicine, however the statistical methods we cover will also have much broader application within modern data science. The ultimate goal will be to learn when and where to correctly apply a given method to real questions about real data. Class time will be devoted to introductory lecture, programming language demonstrations with a common dataset, and open discussions of potential applications, including in-class studio hours to help with homework. Students should be prepared to learn to write code scripts that will manipulate statistical tests and graphical output. However, no background experience with programming is assumed. All software used in the course is open-source and students will be required to set up and run weekly assignments on their own laptop computer or on a computer borrowed from the library or RIT’s computer lab.