105 courses with the subject STSCI, each shown exactly as we captured it from the college's catalog, with every element we hold. Where the wording looks broken, that is our reading of the catalog, not the college's text.
STSCI 1380Data Science for All4
This course provides an introduction to data science using the statistical programming language R. We focus on building skills in inferential thinking and computational thinking, guided by the practical questions we seek to answer from data sets arising in medicine, economics and other social sciences. The course starts with essential R programming principles, and how to use R for data manipulation, visualization, and sampling. These techniques are then used to summarize and visualize real data sets, draw meaningful conclusions from those data, and assess the uncertainty surrounding those conclusions. Throughout the process, students will learn to develop hypotheses about their data, and use simulations and statistical techniques to test these hypotheses. The course also covers how to use the Tidyverse open-source R packages to clean and organize complex data sets, and create high quality graphics for data visualization.
This course will cover essential tools for data collection and analysis in the modern age of big data. Students will learn foundational topics in statistics, including the notions of sampling, data summarization and visualization, and statistical inference. Students will also learn basics of R programming required for exploratory data analysis. In addition, modern techniques for simulation based statistical inference will be covered to complement classical statistical inference.
STSCI 2100Introductory Statistics and Data Science4
Statistics is about understanding the world through data. We are surrounded by data, so there is a lot to understand. Covers data exploration and display, data gathering methods, probability, and statistical inference methods through contingency tables and linear regression. The emphasis is on thinking scientifically, understanding what is commonly done with data (and doing some of it for yourself), and laying a foundation for further study. Students learn to use statistical software and simulation tools to discover fundamental results. They use computers regularly and the course includes both multimedia materials and a software package. This course does not focus on data from any particular discipline, but will use real-world examples from a wide variety of disciplines and current events.
STSCI 2110Statistical Methods for the Social Sciences II4
A second course in statistics that emphasizes applications to the social sciences. Topics include simple linear regression, multiple linear regression (theory, model building, and model diagnostics), and the analysis of variance. Computer packages are used extensively.
The course will cover the basics of R programming for reading and writing data, executing simple operations on data, performing basic analyses, and producing visual graphics. An emphasis will be placed on practical aspects and understanding of fundamental R objects and functions.
This seven week, two-credit class will cover the regression requirements, hypothesis tests, and interpretation of results. Students will learn to identify the data necessary to perform a regression analysis, evaluate the conditions, and apply the statistical tests. Interpretation of overall results will be made. Independent/group projects by each student will be done. These will consist of identifying an issue of interest, a relevant data set, and analysis using the regression methods. Presentation of the results in verbal and written form will be required. Recommended for students who want to develop applied analysis skills.
This course provides an introduction to data analysis and statistical inference illustrated with biological applications. The computer labs will teach graphical analysis and statistical computation using R. Topics include graphical display, populations and sampling, probability distributions, expectation and variance, estimation, testing, correlation, regression, contingency tables, and the design of experiments. Emphasis is on concepts and the careful modeling of biological data, so that statistical methods are applied properly, pitfalls are avoided, and sound conclusions are reached.
Students will be able to perform a variety of basic statistical analyses including: t-tests, two-sample t-tests, tests for categorical data, and linear regression.
Statistics courses usually use clean and well-behaved data, this leaves many unprepared for the messiness and chaos of data in the real world. This course will follow on from STSCI 2120 and cover more advanced data wrangling topics including how to tidy data using the tidyverse R packages to better facilitate data analysis. This includes string processing with regular expressions, manipulating date and time data, web scraping, and text mining. Data visualization topics will cover visualization principles, the use of ggplot2 to create custom plots, and how to communicate data-driven findings.
Statistics courses usually use clean and well-behaved data, this leaves many unprepared for the messiness and chaos of data in the real world. This course aims to prepare students for dealing with data using the R programming language. The introduction will overview the basic R syntax, foundational R programming concepts such as data types, vectors arithmetic, and indexing, and importing data into R from different file formats. The data wrangling topics include how to tidy data using the tidy verse to better facilitate analysis, string processing with regular expressions and with dates and times as file formats, web scraping, and text mining. Data visualization topics will cover visualization principles, the use of ggplot2 to create custom plots, and how to communicate data-driven findings.
This course provides an introduction to probability and parametric inference. Topics include: random variables, standard distributions, the law of large numbers, the central limit theorem, likelihood-based estimation, the method of moments, sampling distributions and confidence intervals.
This course will cover financial mathematics and financial instruments relevant to exam FM offered by the Society of Actuaries. Topics on the present and accumulated value of future cash flows will be covered including the measurement of simple and compound interest, annuities, yield rates, amortization schedules, bonds.
Theory and application of statistical sampling, especially in regard to sample design, cost, estimation of population quantities, and error estimation. Assessment of nonsampling errors. Discussion of applications to social and biological sciences and to business problems.
This course provides an introduction to probability and parametric inference. Topics include: random variables, standard distributions, the law of large numbers, the central limit theorem, likelihood-based estimation, sampling distributions and hypothesis testing.
Applies linear statistical methods to quantitative problems addressed in biological and environmental research. Methods include linear regression, inference, model assumption evaluation, the likelihood approach, matrix formulation, generalized linear models, single-factor and multifactor analysis of variance (ANOVA), and a brief foray into nonlinear modeling. Carries out applied analysis in a statistical computing environment.
STSCI 3510Stochastic Processes for Decision-Making4
Uses basic concepts and techniques of random processes to construct models for a variety of problems of practical interest. Topics include: the Poisson process, Markov chains, renewal theory, models for queuing, and reliability.
Integrated Ethics in Data Science examines your responsibilities in data analysis. Our investigation starts with the aggregated impacts of data science on fairness, privacy, and justice outcomes for groups and individuals. Use of supplied data and applications are analyzed using agency, moral imagination, and virtue ethics. Responsible practices in data science, codes of conduct, and current regulations will be applied. Evaluation of the act of speaking up, supporting others and working with an ethics committee will develop professional skills. Case studies from legal issues, policy concerns, and industry practices provide problems to evaluate the individual choices that led to the results. Course success depends on your frequent written work, engagement in small group discussions, and planning for professional practice.
We start off with a detailed refresher for Linear Regression. We then turn to popular methods for classification including Logistic Regression and Discriminant Analysis. Finally, we consider more advanced topics which may include - depending on the audience - Resampling Methods, Tree-based Methods, or Support Vector Machines. The statistics software R is introduced and used for applications.
Causal claims are essential in both science and policy. Would a new experimental drug improve disease survival? Would a new advertisement cause higher sales? Would a person's income be higher if they finished college? These questions involve counterfactuals: outcomes that would be realized if a treatment were assigned differently. This course will define counterfactuals mathematically, formalize conceptual assumptions that link empirical evidence to causal conclusions, and engage with statistical methods for estimation. Students will enter the course with knowledge of statistical inference: how to assess if a variable is associated with an outcome. Students will emerge from the course with knowledge of causal inference: how to assess whether an intervention to change that input would lead to a change in the outcome.
The focus of this course is the theory and application of the general linear model expressed in its matrix form. Topics will include: least squares estimation, multiple linear regression, coding for categorical predictors, residual diagnostics, anova decomposition, polynomial regression, model selection techniques, random effects and mixed models, maximum likelihood estimation and distributional theory assuming normal errors. Homework assignments will involve computation using the R statistical package.
STSCI 4050Modern Regression Models for Data Science4
This course will cover several modern regression techniques. The topics to be covered include linear regression (review), model selection, regularization methods (ridge and lasso), logistic/probit/Poisson regressions, causal inference, quantile regression, and nonparametric curve estimation. Some topics might be omitted because of time constraints. The course discusses theory and practice of those methods, but is more focused on theory."
STSCI 4060Python Programming and its Applications in Statistics4
The first part of the course teaches basic Python programming knowledge and skills, such as Python variables, data containers, language controls, functions, objects, class, data structures, regular expressions, graphics, GUI, Jupyter notebook, etc. The second part deals with Python application in statistics (e.g., 2D/3D data visualization and statistical analysis, using some important Python packages for statistical computing and machine learning, for example, Numpy, Scipy, Pandas, and Scikit-learn, etc.) Python-database integration (e.g., access, update and control an Oracle database), and Python web services (e.g., database-driven dynamic webpages using Python CGI scripts). These techniques are utilized in a comprehensive course project.
Introduction to classical theory of parametric statistical inference that builds on the material covered in STSCI 3080. Topics include: sampling distributions, principles of data reduction, likelihood, parameter estimation, hypothesis testing, interval estimation, and basic asymptotic theory.
This course is on the basics of multivariate statistical analysis. The focus ison the applied side, and the students will learn by examples of multiple real-life datasets. Studentswill learn to visualize the datasets and conduct simple statistical analysis using linear/nonlinearmethods. We will also cover web-scraping and data cleaning.
Categorical data analysis, including logistic regression, log-linear models, stratified tables, matched pairs analysis, polytomous response, and ordinal data. Applications in biological, biomedical and social sciences.
This course begins with a discussion of some general principles of experimental design. Classical designs are covered in detail, motivated by real data applications. These include completely randomized, randomized block, balanced incomplete block, split-plot, repeated measures and fractional factorial designs. If time permits rank-based nonparametric versions of the classical designs will also be covered.
STSCI 4270Introduction to Survival Analysis and Loss Models3
Develops and uses statistical methods appropriate for analyzing right-censored (i.e., incomplete) time-to-event data. Topics covered include nonparametric estimation (e.g., life table methods, Kaplan Meier estimator), nonparametric methods for comparing the survival experience of two or more populations, and semiparametric and parametric methods of regression for censored outcome data. Emphasis is given to applications in medicine and actuarial studies. Substantial use is made of the R statistical software package.
This course is designed to provide students with an introduction to statistical computing. The class will cover the basics of programming; numerical methods for optimization and linear algebra and their application to statistical estimation, generating random variables, bootstrap, jackknife and permutation methods, Markov Chain Monte Carlo methods, Bayesian inference and computing with latent variables.
Introduces statistical tools for the analysis of time-dependent data. Data analysis and application will be an integral part of this course. Topics include linear, nonlinear, seasonal, multivariate modeling, and financial time series.
The course teaches statistical methods used in modeling risk in asset returns. Students in this course will be able to: identify time series dependency in selected financial data, analyze trade-off between risk and return of a portfolio, analyze tail risk in context of asset returns, and apply factor analysis in context of asset returns.
STSCI 4630Operations Research Tools for Financial Engineering4
Introduction to the applications of OR techniques, e.g., probability, statistics, and optimization, to finance and financial engineering. The course reviews probability and statistics and surveys assets returns, ARIMA time series models, portfolio selection using quadratic programming, regression, CAPM and factor models, option pricing, GARCH models, fixed-income securities, and resampling techniques. Covers the use of R for statistical calculations, simulation, and optimization.
Neural networks form the backbone of modern artificial intelligence methodologies. This course will survey various neural networks architectures with a heavy emphasis on practical application to their specific data use cases. Students will explore how neural networks generalize classical statistical models and function estimation techniques, and how statistical principles inform model design, optimization, and evaluation. Topics include feedforward architectures, stochastic gradient descent, regularization and model selection, convolutional and recurrent networks, and an introduction to attention-based models.
Statistical learning methods based on decision trees are flexible and intuitive solutions to classification and regression problems, which nevertheless are straightforward to interpret. This course will introduce the classification and regression tree model (CART), as well as more advanced tree-based approaches, including, as time permits, gradient-boosted trees, bagged trees, random forest, and Bayesian additive regression trees. We will use the R Programming Language to apply these methods to real data and compare them to other regression and classification methods. We will discuss advantages and disadvantages of tree-based methods, and cover presentation, interpretation, and visualization of results.
The goal of this course is to teach you why machine learning works and how to implement it. We will cover the essentials of learning theory, including the probably approximately correct (PAC) framework and the bias-complexity tradeoff. We will then see how these concepts shed light on the mathematics behind linear regression, logistic regression, boosting (and AdaBoost), support vector machines and neural networks. We cover clustering algorithms and how to implement them. Data will be analyzed using modern software packages with the above algorithms, with the aim of reinforcing the mathematics behind them.
STSCI 4780Bayesian Data Analysis: Principles and Practice4
Bayesian data analysis uses probability theory as a kind of calculus of inference, specifying how to quantify and propagate uncertainty in data-based chains of reasoning. Students will learn the fundamental principles of Bayesian data analysis, and how to apply them to varied data analysis problems across science and engineering. Topics include: basic probability theory, Bayes's theorem, linear and nonlinear models, hierarchical and graphical models, basic decision theory, and experimental design. There will be a strong computational component, using a high-level language such as R or Python, and a probabilistic language such as BUGS or Stan.
In this course students will learn about the consulting process using data science. They will understand how statistics and data science knowledge can be applied to real world questions starting with exploratory analysis and then applying specific modeling tools. In addition to analyzing a problem empirically, they will also learn core professional soft skills required to work with a client. These skills include designing and implementing a work plan, learning to communicate efficiently, presentation of a final deliverable to the client, as well as learning good practices to conduct data analysis, documentation, quality control and collaboration. Through hands-on experience and real-world examples, students will develop a basic understanding of consulting, and be familiar with the professional standard expected in the industry.
This course introduces statistical thinking and data analysis in the context of sports. Students learn to explore, visualize, and model sports data to evaluate performance, predict outcomes, and support strategic decisions. Emphasis is placed on the distinctive features of sports data, including player metrics, event outcomes, and temporal or spatial patterns, as well as on applying statistical and computational tools to real-world problems. Topics include exploratory data analysis, simulation, confidence intervals, hypothesis testing, resampling methods, regression, and introductory machine learning techniques such as decision trees, random forests, and clustering. By the end of the course, students will be able to interpret models, assess performance, and communicate data-driven insights in the sports domain. Students are expected to have proficiency in at least one computer language or software package capable of statistical analysis (e.g., R, Python, Stata, or MATLAB). A working understanding of basic probability, statistics, and a familiarity with linear regression, properties of the normal distribution, and common types of errors.
This course will give students the opportunity to apply the statistical knowledge gained in their courses to real-life problems. Students will be integrated in the Cornell Statistical Consulting Unit (CSCU) and be exposed to various areas in which statistical methods are applied. Students will gain experience in choosing appropriate statistical procedures and their implementations in various statistical software packages. They will also learn how to communicate effectively to understand the client's problem and to explain methods and results to non-statisticians.
Students assist in teaching a course appropriate to their previous training. Students meet with a discussion or laboratory section and regularly discuss objectives with the course instructor.
Problem solving sessions to prepare students for the first four actuarial examinations (probability, financial mathematics, statistical modeling, and risk management).
This practicum course provides an applied framework for understanding and implementing data analytics within the sport industry. Students will engage in individual or collaborative projects addressing authentic analytical challenges in sport contexts. Through these projects, students will integrate analytical knowledge with practical application, critically examining how quantitative methods and data-driven approaches inform strategic and evidence-based decision-making.
The capstone course is designed as a culmination of students’ sports education at Cornell as they work on group projects tackling pressing industry problems. Students can work on projects provided by the course instructor or on individual (and instructor approved projects). Projects will be jointly supervised by industry partners and Cornell faculty. In addition to written deliverables students will present their projects and insights to faculty and industry advisors.
This speaker series will feature prominent voices from across the sports industry, including commissioners of professional leagues, team executives, and attorneys specializing in sports mergers and acquisitions. The series is expected to include 12 to 14 speakers, meeting on a weekly basis. This course is designed to exposure students to pressing issues facing industry executives in today’s fast paced sports environment.
STSCI 5010Applied Statistical Computation with SAS4
This course teaches the basics of SAS (Statistical Analysis System) programming and the SAS Enterprise Miner software. This course is composed two modules. The first module, in the first 12 weeks, covers the objectives tested on the SAS Base Programming for SAS 9 Exam, including basic SAS programming concepts, producing reports, creating and modifying SAS data sets, reading various types of raw data and other data handling techniques. At the end of module 1, all the students will take the SAS Base Programming for SAS 9 Exam, which is administered by the MPS Program in Applied Statistics on the Cornell campus in conjunction with the SAS Institute, Inc. The second module, in the last three weeks, introduces the SAS Enterprise Miner software and cluster analysis. Students will learn how to use the SAS Enterprise Miner software and SAS procedures to do cluster analysis. Students should have taken a basic statistics course prior to enrolling in this course.
The focus of this course is the theory and application of the general linear model expressed in its matrix form. Topics will include: least squares estimation, multiple linear regression, coding for categorical predictors, residual diagnostics, anova decomposition, polynomial regression, model selection techniques, random effects and mixed models, maximum likelihood estimation and distributional theory assuming normal errors. Homework assignments will involve computation using the R statistical package.
Statistics courses usually use clean and well-behaved data, this leaves many unprepared for the messiness and chaos of data in the real world. This course aims to prepare students for dealing with data using the R programming language. The introduction will overview the basic R syntax, foundational R programming concepts such as data types, vectors arithmetic, and indexing, and importing data into R from different file formats. The data wrangling topics include how to tidy data using the tidy verse to better facilitate analysis, string processing with regular expressions and with dates and times as file formats, web scraping, and text mining. Data visualization topics will cover visualization principles, the use of ggplot2 to create custom plots, and how to communicate data-driven findings.
STSCI 5045Python Programming and its Applications in Statistics4
The first part of the course teaches basic Python programming knowledge and skills, such as Python variables, data containers, language controls, functions, objects, class, data structures, regular expressions, graphics, GUI, Jupyter notebook, etc. The second part deals with Python application in statistics (e.g., 2D/3D data visualization and statistical analysis, using some important Python packages for statistical computing and machine learning, for example, Numpy, Scipy, Pandas, and Scikit-learn, etc.) Python-database integration (e.g., access, update and control an Oracle database), and Python web services (e.g., database-driven dynamic webpages using Python CGI scripts). These techniques are utilized in a comprehensive course project. (MHA-EL)
STSCI 5050Modern Regression Models for Data Science4
This course will cover several modern regression techniques. The topics to be covered include linear regression (review), model selection, regularization methods (ridge and lasso), logistic/probit/Poisson regressions, causal inference, quantile regression, and nonparametric curve estimation. Some topics might be omitted because of time constraints. The course discusses theory and practice of those methods, but is more focused on theory.
STSCI 5060Database Management and SAS High Performance Computing with DBMS4
Using relational databases in statistical computing has become more and more important. The knowledge and skill of database management and the ability to combine this knowledge and skill with statistical analysis software tools, such as SAS, are a critical qualification of a statistical analyst. In this course we will study 1) the basics of modern relational database management systems, including database analysis, design and implementation, 2) database application in advanced SAS programming and, 3) SAS high performance computing using database-related techniques.
Concepts, challenges, and industry trends of big data, with a focus on the Hadoop system. Topics include: basics of the Apache Hadoop platform and Hadoop ecosystem; the Hadoop distributed file system (HDFS); MapReduce or its alternative, a parallel programming model for distributed processing of large data sets; common big data tools, such as Pig (a procedural data processing language for Hadoop parallel computation), Hive (a declarative SQL-like language to handle Hadoop jobs), HBase (the most popular NoSQL database), and YARN; case studies; and integration of Hadoop with statistical software packages, e.g., SAS and R.
This course provides an introduction to probability and parametric inference. Topics include: random variables, standard distributions, the law of large numbers, the central limit theorem, likelihood-based estimation, sampling distributions and hypothesis testing, as well as an introduction to Bayesian methods. Some assignments may involve computation using the R programming language.
Introduction to classical theory of parametric statistical inference that builds on the material covered in STSCI 3080. Topics include sampling distributions, principles of data reduction, likelihood, parameter estimation, hypothesis testing, interval estimation, and basic asymptotic theory.
Theory and application of statistical sampling, especially in regard to sample design, cost, estimation of population quantities, and error estimation. Assessment of nonsampling errors. Discussion of applications to social and biological sciences and to business problems.
STSCI 5110Statistical Methods for the Social Sciences II4
Second course in statistics that emphasizes applications to the social sciences. Topics include simple linear regression, multiple linear regression (theory, model building, and model diagnostics), and the analysis of variance. Computer packages are used extensively.
This course is on the basics of multivariate statistical analysis. The focus ison the applied side, and the students will learn by examples of multiple real-life datasets. Studentswill learn to visualize the datasets and conduct simple statistical analysis using linear/nonlinearmethods. We will also cover web-scraping and data cleaning.
The course will cover the basics of R programming for reading and writing data, executing simple operations on data, performing basic analyses, and producing visual graphics. An emphasis will be placed on practical aspects and understanding of fundamental R objects and functions.
This course begins with a discussion of some general principles of experimental design. Classical designs are covered in detail, motivated by real data applications. These include completely randomized, randomized block, balanced incomplete block, split-plot, repeated measures and fractional factorial designs. If time permits rank-based nonparametric versions of the classical designs will also be covered.
This course provides an introduction to data analysis and statistical inference illustrated with biological applications. The computer labs will teach graphical analysis and statistical computation using R. Topics include graphical display, populations and sampling, probability distributions, expectation and variance, estimation, testing, correlation, regression, contingency tables, and the design of experiments. Emphasis is on concepts and the careful modeling of biological data, so that statistical methods are applied properly, pitfalls are avoided, and sound conclusions are reached.
Categorical data analysis, including logistic regression, log-linear models, stratified tables, matched pairs analysis, polytomous response, and ordinal data. Applications in biological, biomedical and social sciences.
Students will be able to perform a variety of basic statistical analyses including: t-tests, two-sample t-tests, tests for categorical data, and linear regression.
Applies linear statistical methods to quantitative problems addressed in biological and environmental research. Methods include linear regression, inference, model assumption evaluation, the likelihood approach, matrix formulation, generalized linear models, single-factor and multifactor analysis of variance (ANOVA), and a brief foray into nonlinear modeling. Carries out applied analysis in a statistical computing environment.
Statistics courses usually use clean and well-behaved data, this leaves many unprepared for the messiness and chaos of data in the real world. This course will follow on from STSCI 5120 and cover more advanced data wrangling topics including how to tidy data using the tidyverse R packages to better facilitate data analysis. This includes string processing with regular expressions, manipulating date and time data, web scraping, and text mining. Data visualization topics will cover visualization principles, the use of ggplot2 to create custom plots, and how to communicate data-driven findings.
STSCI 5270Introduction to Survival Analysis and Loss Models3
Develops and uses statistical methods appropriate for analyzing right-censored (i.e., incomplete) time-to-event data. Topics covered include nonparametric estimation (e.g., life table methods, Kaplan Meier estimator), nonparametric methods for comparing the survival experience of two or more populations, and semiparametric and parametric methods of regression for censored outcome data. Emphasis is given to applications in medicine and actuarial studies. Substantial use is made of the R statistical software package.
This course is designed to provide students with an introduction to statistical computing. The class will cover the basics of programming; numerical methods for optimization and linear algebra and their application to statistical estimation, generating random variables, bootstrap, jackknife and permutation methods, Markov Chain Monte Carlo methods, Bayesian inference and computing with latent variables.
Introduces statistical tools for the analysis of time-dependent data. Data analysis and application will be an integral part of this course. Topics include linear, nonlinear, seasonal, multivariate modeling, and financial time series.
Integrated Ethics in Data Science examines your responsibilities in data analysis. Our investigation starts with the aggregated impacts of data science on fairness, privacy, and justice outcomes for groups and individuals. Use of supplied data and applications are analyzed using agency, moral imagination, and virtue ethics. Responsible practices in data science, codes of conduct, and current regulations will be applied. Evaluation of the act of speaking up, supporting others and working with an ethics committee will develop professional skills. Case studies from legal issues, policy concerns, and industry practices provide problems to evaluate the individual choices that led to the results. Course success depends on your frequent written work, engagement in small group discussions, and planning for professional practice.
The course teaches statistical methods used in modeling risk in asset returns. Students in this course will be able to: identify time series dependency in selected financial data, analyze trade-off between risk and return of a portfolio, analyze tail risk in context of asset returns, and apply factor analysis in context of asset returns.
STSCI 5630Operations Research Tools for Financial Engineering4
Introduction to the applications of OR techniques, e.g., probability, statistics, and optimization, to finance and financial engineering. The course reviews probability and statistics and surveys assets returns, ARIMA time series models, portfolio selection using quadratic programming, regression, CAPM and factor models, option pricing, GARCH models, fixed-income securities, and resampling techniques. Covers the use of R for statistical calculations, simulation, and optimization.
Regression, ARIMA, GARCH, stochastic volatility, and factor models. Calibration of financial engineering models, estimation of diffusion models, estimation of risk measures, multivariate models and copulas, bayesian statistics. Students are instructed in the use of R software.
Neural networks form the backbone of modern artificial intelligence methodologies. This course will survey various neural networks architectures with a heavy emphasis on practical application to their specific data use cases. Students will explore how neural networks generalize classical statistical models and function estimation techniques, and how statistical principles inform model design, optimization, and evaluation. Topics include feedforward architectures, stochastic gradient descent, regularization and model selection, convolutional and recurrent networks, and an introduction to attention-based models.
Statistical learning methods based on decision trees are flexible and intuitive solutions to classification and regression problems, which nevertheless are straightforward to interpret. This course will introduce the classification and regression tree model (CART), as well as more advanced tree-based approaches, including, as time permits, gradient-boosted trees, bagged trees, random forest, and Bayesian additive regression trees. We will use the R Programming Language to apply these methods to real data and compare them to other regression and classification methods. We will discuss advantages and disadvantages of tree-based methods, and cover presentation, interpretation, and visualization of results.
We start off with a detailed refresher for Linear Regression. We then turn to popular methods for classification including Logistic Regression and Discriminant Analysis. Finally, we consider more advanced topics which may include - depending on the audience - Resampling Methods, Tree-based Methods, or Support Vector Machines. The statistics software R is introduced and used for applications.
The goal of this course is to teach you why machine learning works and how to implement it. We will cover the essentials of learning theory, including the probably approximately correct (PAC) framework and the bias-complexity tradeoff. We will then see how these concepts shed light on the mathematics behind linear regression, logistic regression, boosting (and AdaBoost), support vector machines and neural networks. We cover clustering algorithms and how to implement them. Data will be analyzed using modern software packages with the above algorithms, with the aim of reinforcing the mathematics behind them.
STSCI 5780Bayesian Data Analysis: Principles and Practice4
Bayesian data analysis uses probability theory as a kind of calculus of inference, specifying how to quantify and propagate uncertainty in data-based chains of reasoning. Students will learn the fundamental principles of Bayesian data analysis, and how to apply them to varied data analysis problems across science and engineering. Topics include: basic probability theory, Bayes's theorem, linear and nonlinear models, hierarchical and graphical models, basic decision theory, and experimental design. There will be a strong computational component, using a high-level language such as R or Python, and a probabilistic language such as BUGS or Stan.
This course introduces statistical thinking and data analysis in the context of sports. Students learn to explore, visualize, and model sports data to evaluate performance, predict outcomes, and support strategic decisions. Emphasis is placed on the distinctive features of sports data, including player metrics, event outcomes, and temporal or spatial patterns, as well as on applying statistical and computational tools to real-world problems. Topics include exploratory data analysis, simulation, confidence intervals, hypothesis testing, resampling methods, regression, and introductory machine learning techniques such as decision trees, random forests, and clustering. By the end of the course, students will be able to interpret models, assess performance, and communicate data-driven insights in the sports domain. Students are expected to have proficiency in at least one computer language or software package capable of statistical analysis (e.g., R, Python, Stata, or MATLAB). A working understanding of basic probability, statistics, and a familiarity with linear regression, properties of the normal distribution, and common types of errors.
This course will focus on specific career and professional development topics for Master of Professional Studies in Applied Statistics students. Through lectures, workshops and seminars from MPS Statistics and Cornell Career Services staff and alumni, you will learn job search strategies, effective resume writing, professional dining etiquette, the art of effective presentation, as well as the opportunity for you to choose a minimum of three workshops to further develop career related strategies according to your personal needs.
STSCI 5954Project Development and Professional Communication2
Students will learn core professional skills required to work with a client. These skills include designing and implementing a work plan, learning to communicate efficiently, presentation of a final deliverable to the client, as well as learning good practices to conduct data analysis, documentation, quality control and collaboration. Through hands-on experience and real-world examples, students will develop a basic understanding of consulting, and be familiar with the professional standard expected in the industry.
This course is designed to run in parallel to STSCI 5999 (Applied Statistics MPS Data Analysis Project) and is a required part of the final-semester client project for MPS students. In this course students from each project team meet at regular intervals and briefly present their project milestones such as initiation of the project, data cleaning, analysis, etc. as well as raise any concern regarding any roadblock they are facing.
Problem solving sessions to prepare students for the first four actuarial examinations (probability, financial mathematics, statistical modeling, and risk management).
This practicum course provides an applied framework for understanding and implementing data analytics within the sport industry. Students will engage in individual or collaborative projects addressing authentic analytical challenges in sport contexts. Through these projects, students will integrate analytical knowledge with practical application, critically examining how quantitative methods and data-driven approaches inform strategic and evidence-based decision-making.
This speaker series will feature prominent voices from across the sports industry, including commissioners of professional leagues, team executives, and attorneys specializing in sports mergers and acquisitions. The series is expected to include 12 to 14 speakers, meeting on a weekly basis. This course is designed to exposure students to pressing issues facing industry executives in today’s fast paced sports environment.
STSCI 5990Directed Studies in Applied Statistics1-4
For individual or group research projects conducted under the direction of a member of STSCI faculty or instructors in a special area of statistical science that is not covered by regular course offerings.
STSCI 5999Applied Statistics MPS Data Analysis Project4
It is a long-term, in-depth statistical analysis of real-world dataset using various statistical methods and computer packages (such as SAS, R, SPSS, etc.). Students work in teams to solve business, managerial or scientific problems for clients. Projects are assigned by Department of Statistics and Data Science to teams. Each team has a dedicated faculty adviser, who supervises the projects and assigns grades. Grades are assigned individually to team members. In special cases teams may collaborate with groups in other Departments, e.g., M.Eng. project teams.
Develops and uses statistical methods to analyze data arising from a wide variety of applications. Topics include descriptive statistics, point and interval estimation, hypothesis testing, inference for a single population, comparisons between two populations, one- and two-way analysis of variance, comparisons among population means, analysis of categorical data, and correlation and regression analysis. Introduces interactive computing through statistical software. Emphasizes basic principles and criteria for selection of statistical techniques.
Continuation of STSCI 6010. Emphasizes the use of multiple regression analysis, analysis of variance, and related techniques to analyze data in a variety of situations. Topics include an introduction to data collection techniques; least squares estimation; multiple regression; model selection techniques; detection of influential points, goodness-of-fit criteria; principles of experimental design; analysis of variance for a number of designs, including multi-way factorial, nested, and split plot designs; comparing two or more regression lines; and analysis of covariance. Emphasizes appropriate design of studies before data collection, and the appropriate application and interpretation of statistical techniques. Practical applications are implemented using a modern, widely available statistical package.
Modern statistical methods routinely used in practice for analyzing large, complex datasets require intensive computation. This course covers topics in statistical computing, including numerical optimization and finding zeros (likelihood and related techniques), random number generation and Monte Carlo methods, bootstrap and subsampling, dimension reduction, nonlinear predictive modeling.
STSCI 6610Spatial and Spatial-Temporal Data Analysis3
The course covers the analysis of spatial and/or spatial-temporal data, defined as data that include references to spatial locations or regions. Topics may include spatial interpolation, regression with spatial data, Gaussian process models, construction of covariance functions, computational techniques for large datasets, hierarchical non-Gaussian models, analysis of point patterns, forecasting, and Bayesian methods.
This class will cover fundamental concepts in mathematical statistics, including both finite sample and asymptotic theory. Specific topics include: elements of risk optimality, Cramer-Rao-type bounds; M-estimation with an emphasis on Maximum Likelihood Estimation, asymptotic efficiency, asymptotic testing under fixed and local alternatives; multiple testing under FDR control; estimation in high dimensions and adaptation to sparsity, the analysis of Lasso-type estimators; elements of concentration inequalities.
Focuses on the foundations of statistical inference, with an emphasis on asymptotic methods and the minimax optimality criterion. In the first part, the solution of the classical problem of justifying Fisher's information bound in regular statistical models will be presented. This solution will be obtained applying the concepts of contiguity, local asymptotic normality and asymptotic minimaxity. The second part will be devoted to nonparametric estimation, taking a Gaussian regression model as a paradigmatic example. Key topics are kernel estimation and local polynomial approximation, optimal rates of convergence at a point and in global norms, and adaptive estimation. Optional topics may include irregular statistical models, estimation of functionals and nonparametric hypothesis testing.
This course will discuss several probabilistic results and techniques that are useful in mathematical statistics. Limit theorems and stochastic inequalities are two key components of the course. Topics to be covered include complements of basic limit theorems (Skorohod representation theorem, Berry-Esseen bound, Delta method), empirical distribution and quantile functions, U-statistics, bootstrap (and subsampling if time permits), stochastic convergence in metric spaces, and some elements of modern empirical process theory.
Priors, posteriors, Bayes estimators, decision theory, asymptotic theory, Bayes factors, credible regions, hierarchical models, nonparametric Bayes, computational methods, Bayesian robustness, and applications. This is a theory-oriented course intended for PhD students. For a more applied introduction to Bayesian statistics, students should take STSCI 4780, Bayesian Data Analysis: Principles and Practice, which is offered in the spring.
Learning theory is an important branch of modern statistics. This course gives an overview of various topics and proof techniques that include concentration inequalities, Bayes rules, reject option, margin condition, local averaging methods, universal consistency, empirical risk minimization, convex surrogate losses, Rademacher complexity, VC theory, structural risk minimization, sparse methods, low-rank regression, topic models, latent factor models and interpolation methods.
STSCI 6940Graduate Special Topics in Statistics1-4
Topics are arranged at the beginning of the semester for individual study or for group discussions. Or, students may elect to undertake a project in statistics. The work is supervised by a professor in this subject area.
Properties of the multivariate normal distribution. Distribution theory for quadratic forms. Properties of least squares and maximum likelihood estimates. Methods for fixed-effect models of less than full rank. Analysis of balanced and unbalanced mixed-effects models. Restricted maximum likelihood estimation. Some use of software packages and illustrative examples.
A theoretical development of generalized linear models and related topics including categorical data problems, generalized additive models, and generalized linear mixed models.
By interacting with consultants from the Cornell Statistical Consulting Unit (CSCU) and their clients, students will learn to provide statistical consulting to researchers from a variety of disciplines. Students will gain experience in choosing and implementing appropriate statistical procedures as well as communicating effectively to understand the client's problem and to explain methodology and results to non-statisticians.
This course focuses on practical data analysis involving advanced statistical methods applied to real datasets. Students will be required to attend meetings with Cornell Statistical Consulting Unit (CSCU) clients. Emphasis will be placed on report writing, communicating effectively and explaining methodology and results to non-statisticians. Project lengths will vary based on the scope of the problems and could range from several short projects to a single semester-long project. Students will be required to collaborate in teams on some projects.
