Assistant Professor of Statistics
Ethan’s research focuses on creating methods for analyzing high dimensional neural signals in order to understand the neural processing that underlies object recognition, working memory and other cognitive processes.
Ethan’s teaching interest span a range of topics including statistics, machine learning, data science and computational neuroscience.
This class is an introduction to statistical methods that are useful for analyzing data. Topics will include descriptive statistics (summary statistics and graphical methods), and resampling and parameter inference methods for calculating confidence intervals and conducting hypothesis tests. Students will learn to use the R programming language to explore statistical concepts and to analyze real data. Assignments will consist of weekly problem sets that cover newly introduced topics, and cumulative learning checks that reinforce the topics that have been covered. By the end of the class students should be able to understand the principles that underlie statistical analyses used in a variety of fields, and should be able to apply statistical methods to gain insight into data that they collect.
The rise of computers and large datasets over the past 30 years has led to the development of new methods for analyzing data. These 'statistical learning' methods blend classical statistical concepts with ideas from computer science and are widely used by data scientists to analyze complex datasets. In this class we will cover the basic concepts in statistical learning including: regression, supervised learning (classification), unsupervised learning (clustering and dimensionality reduction), cross-validation methods, and model selection. We will use the R programming language to explore the usefulness of different methods and to analyze real data. The class work will consist of weekly programming problems and a final project. Prerequisites: Prior experience with programming and statistics, either through a class or from other experiences.
Statistics is a field that tries to interpret data in the face of random variation. The methods used in statistics are often abstract which can make them hard to understand. Baseball is a simple game that contains a high degree of randomness, and thus offers a great way to ground statistical concepts in terms of simple actions taken by the players. In this class we examine key concepts in statistics using baseball as a motivating example for how to answer concrete questions in the face of noisy data. We will also discuss how newer statistics (known as sabermetrics) have been used to gain additional insights, and we will relate these ideas to other sports. Assignments will consist of weekly problem sets and a short final project. By taking this class, students will develop an understanding of key statistical concepts that will be useful for interpreting data from many fields.
The activity in our brains allows us to perform complex behaviors and (presumably) gives rise to our conscious experience. A variety of technologies exist to record neural activity at different spatial and temporal scales. However, in order to turn these recorded signals into meaningful insights about how the brain works, statistical methods are needed. In this course we will discuss several statistical analyses that are used to analyze neural data. The types of data we will examine include electro/magneto encephalographic signals (EEG/MEG), functional magnetic resonance imaging responses (fMRI) and neural spiking activity. The methods covered will range from classical univariate statistics such as ANOVAs, to multivariate machine-learning-based 'decoding' analyses. Exercises will consist of analyzing real data from these different modalities, and there will be a final project where one dataset is explored in more detail. Prerequisites: completed courses equivalent to Introduction to Statistics and Introduction to Computer Programming.
This class will discuss the successes and failures of different fields to make accurate predictions. Areas we will cover include: politics, weather forecasting, economics, sports, seismology, games, and climate science. We will use Nate Silver's book 'The signal and the noise' along with primary research and news articles. Assignments will include writing short weekly summaries on current topics in prediction, giving in class presentations, and completing a final project. By the end of the class students should have an understanding of what makes prediction problems difficult and what are some of the cutting edge research problems in predictive analytics.