2024-2025 Catalog

DATA 11800. Introduction to Data Science I. 100 Units.

Data science provides tools for gaining insight into specific problems using data, through computation, statistics and visualization. This course introduces students to all aspects of a data analysis process, from posing questions, designing data collection strategies, management+storing and processing of data, exploratory tools and visualization, statistical inference, prediction, interpretation and communication of results. Simple techniques for data analysis are used to illustrate both effective and fallacious uses of data science tools. Although this course is designed to be at the level of mathematical sciences courses in the Core, with little background required, we expect the students to develop computational skills that will allow them to analyze data. Computation will be done using Python and Jupyter Notebook.

Instructor(s): S. Lange; A. Kube; W. Trimble ; M. Banuelos; K. Smalenberger;     Terms Offered: Autumn Spring Winter
Prerequisite(s): None
Equivalent Course(s): STAT 11800

DATA 11900. Introduction to Data Science II. 100 Units.

This course is the second quarter of a two-quarter systematic introduction to the foundations of data science, as well as to practical considerations in data analysis. A broad background on probability and statistical methodology will be provided. More advanced topics on data privacy and ethics, reproducibility in science, data encryption, and basic machine learning will be introduced. We will explore these concepts with real-world problems from different domains.

Instructor(s): D. Biron; B. Trok; A. Nussbaum; A. Kube     Terms Offered: Autumn Spring Winter
Prerequisite(s): DATA 11800 or consent of instructor.
Equivalent Course(s): STAT 11900

DATA 12000. Computer Science for Data Science. 100 Units.

This course teaches computational thinking and programming skills to students in the Data Science program. Topics include control structures and basic data types, abstraction and functional decomposition, classes and objects in Python, basic algorithms, and an introduction to computer structure and low level representation of data types. Examples will include the application of tools such as scraping web pages and rudimentary machine learning to a variety of fields.

Instructor(s): W. Trimble, A. Nussbaum     Terms Offered: Spring Winter
Prerequisite(s): DATA 11800

DATA 13600. Introduction to Data Engineering. 100 Units.

Data-driven models are revolutionizing science and industry. Scalable systems are needed to collect, stream, process, and validate data at scale. This course is an introduction to "big" data engineering where students will receive hands-on experience building and deploying realistic data-intensive systems. It will cover streaming, data cleaning, relational data modeling and SQL, and Machine Learning model training. A core theme of the course is "scale," and we will discuss the theory and the practice of programming with large external datasets that cannot fit in main memory on a single machine. The course will consist of bi-weekly programming assignments, a midterm examination, and a final.

Instructor(s): W. Trimble     Terms Offered: Autumn Spring
Prerequisite(s): DATA 11900, DATA 12000 or CMSC 12200 or CMSC 16200

DATA 13820. Data Science in Quantitative Finance and Risk Management. 100 Units.

Have you started or are about to start your investment journey? Do you want to know more about terms like "recession" and "volatility," and how they might affect your own bank account? Are you interested in mathematics and its application to human emotions? This course introduces the leading statistical models and methods which financial data researchers use to understand ever-evolving markets and build insightful financial strategies, such as machine learning, risk calculation, and portfolio management . At first, students will learn about the theoretical and applied foundations of regression and classification designs for predicting market patterns. Next, students will gain exposure to proprietary metrics such as Value-at-Risk(VaR) used to evaluate returns/losses of both single and multi-asset portfolios. Lastly, they will experiment with portfolio allocation tactics by visualizing risk-to-reward graphs under various buying and selling conditions. These techniques can be applied to the U.S. and foreign asset classes, including equities, commodities, and cryptocurrencies. Students will experience how professionals in quantitative trading, hedge funds, and risk analytics collaborate to pitch asset strategies to their clients, and form research teams to play a stock market game using the skills they learned throughout the course with the objective of maximizing the teams' portfolio returns. All implementations will be done using Python.

Terms Offered: Summer
Equivalent Course(s): STAT 13820

DATA 20519. Spatial Cluster Analysis. 100 Units.

This course provides an overview of methods to identify interesting patterns in geographic data, so-called spatial clusters. Cluster concepts come in many different forms and can generally be differentiated between the search for interesting locations and the grouping of similar locations. The first category consists of the identification of extreme concentrations of locations (events), such as hot spots of crime events, and the location of geographical concentrations of observations with similar values for one or more variables, such as areas with elevated disease incidence. The second group consists of the combination of spatial observations into larger (aggregate) areas such that internal similarity is maximized (regionalization). The methods covered come from the fields of spatial statistics as well as machine learning (unsupervised learning) and operations research. Topics include point pattern analysis, spatial scan statistics, local spatial autocorrelation, dimension reduction, as well as spatially explicit hierarchical, agglomerative and density-based clustering. Applications range from criminology and public health to politics and marketing. An important aspect of the course is the analysis of actual data sets by means of open source software, such as GeoDa, R or Python.

Instructor(s): L. Anselin and P. Amaral     Terms Offered: Winter
Prerequisite(s): STAT 22000 or equivalent; SOCI 20253/30253 (or equivalent) Introduction to Spatial Data Science required.
Equivalent Course(s): GISC 20519, ENST 20519, MACS 30519, SOCI 20519, SOCI 30519, MACS 20519, GISC 30519

DATA 20559. Spatial Regression Analysis. 100 Units.

This course covers statistical and econometric methods specifically geared to the problems of spatial dependence and spatial heterogeneity in cross-sectional data. The main objective for the course is to gain insight into the scope of spatial regression methods, to be able to apply them in an empirical setting, and to properly interpret the results of spatial regression analysis. While the focus is on spatial aspects, the types of methods covered have general validity in statistical practice. The course covers the specification of spatial regression models in order to incorporate spatial dependence and spatial heterogeneity, as well as different estimation methods and specification tests to detect the presence of spatial autocorrelation and spatial heterogeneity. Special attention is paid to the application to spatial models of generic statistical paradigms, such as Maximum Likelihood and Generalized Methods of Moments. An import aspect of the course is the application of open source software tools such as various R packages, GeoDa and the Python Package PySal to solve empirical problems.

Instructor(s): L. Anselin     Terms Offered: Autumn
Prerequisite(s): An intermediate course in multivariate regression or econometrics. Familiarity with matrix algebra
Equivalent Course(s): SOCI 30559, SOCI 20559, GISC 20559, GISC 30559

DATA 20595. Topics in Spatial Regression Analysis. 100 Units.

This course covers methodological issues that affect the specification and estimation of spatial regression models. The course is organized as a seminar, with a combination of brief lectures, discussion of recent article and lab exercises. Topics will vary. Examples are spatial specification search, spatial effects in models for discrete dependent variables, spatial effects in count models, semi-parametric spatial models, spatial panel data models, spatial treatment effect analysis, spatial interaction models, endogenous regimes, regularization in spatial models, spatial feature engineering, and endogenous spatial weights. An important aspect of the course is the application of open source software tools, specifically those contained in the Python package PySAL.

Instructor(s): L. Anselin     Terms Offered: Winter
Prerequisite(s): An intermediate course in multivariate regression or econometrics. Familiarity with metrix algebra; SOCi 20559/30559 or equivalent is desired, but not required
Equivalent Course(s): SOCI 20595, SOCI 30595

DATA 21100. Mathematical Methods for Data Science I. 100 Units.

This course introduces topics in probability and calculus (III) for data science students. Topics covered include random variables and probability distributions, independence, conditional probability, expected values, the central limit theorem, the definite integral, integration of functions of several variables, sums and transformations of random variables, multivariable calculus - partial derivatives, gradients, and gradient descent.

Instructor(s): S. Lange; S. Bhowmick     Terms Offered: Autumn Winter
Prerequisite(s): MATH 13200 or 15200 or 16200 (could be taken concurrently), DATA 11800, DATA 11900 (could be taken concurrently) or CMSC 12100 or CMSC 14100

DATA 21200. Mathematical Methods for Data Science II. 100 Units.

This course introduces topics in linear algebra for data science students. Topics covered include Vectors spaces, linear transformations and associated subspaces, orthogonality and projections, orthogonal (orthonormal) matrices, eigenvectors and eigenvalues, diagonalization, Det and Tr, matrix decompositions (LU decomposition, singular value decomposition), and transfer matrices and discrete stochastic processes. Assignments and examples focus on results and algorithms that are relevant to machine learning and data science.

Instructor(s): A. Nussbaum, W. Trimble     Terms Offered: Spring Winter
Prerequisite(s): MATH 13200 or 15200 or 16200 (could be taken concurrently), DATA 11800, DATA 11900 (could be taken concurrently) or CMSC 12100 or CMSC 14100

DATA 21300. Models in Data Science. 100 Units.

This course introduces fundamental aspects of modeling data such as regression, linearization and linear models, discrete dynamical systems, Markov chains, continuous dynamical systems, and stability analysis.

Instructor(s): A. Nussbaum     Terms Offered: Autumn Spring
Prerequisite(s): DATA 21100 (or equivalent; could be taken concurrently), DATA 21200 or STAT 24300 or MATH 19620

DATA 22100. Introduction to Machine Learning: Concepts and Applications. 100 Units.

This course introduces topics in current applications of machine learning for Data Science minor students. Topics include machine learning models, supervised and unsupervised learning, loss functions, risk, empirical risk and overfitting, regression and classification, clustering, gradient boosting, decision trees and random forests, and a brief introduction to Neural Networks and deep learning.

Instructor(s): A. Kube     Terms Offered: Spring Winter
Prerequisite(s): DATA 11900 (could be taken concurrently) or CMSC 12200 or CMSC 14200

DATA 22700. Data Visualization and Communication. 100 Units.

This course introduces Data Science minors to best practices for presenting and communicating quantitative data. Principles of data visualization include the use of colors and negative spaces, drawing attention to important details, repetition of design motifs, appropriately using figures and tables, and combining different scales in a single figure. The course also discusses how to avoid common distortions resulting in misleading plots and figures and how to effectively communicate findings. Examples are chosen from a variety of fields, such as the biological sciences, the social sciences, and the media.

Instructor(s): D. Biron, A. Nussbaum     Terms Offered: Autumn
Prerequisite(s): DATA 11800 or CMSC 12100 or CMSC 14100

DATA 23100. Machine Learning Fundamentals: Theory and Practice. 100 Units.

This course introduces topics in current applications of machine learning for Data Science major students. Topics include machine learning models, supervised and unsupervised learning, loss functions, risk, empirical risk and overfitting, regression and classification, clustering, gradient boosting, decision trees and random forests, and a brief introduction to Neural Networks and deep learning.

Instructor(s): A. Schein     Terms Offered: Winter
Prerequisite(s): DATA 12000 (could be taken concurrently) or CMSC 12200 or CMSC 14200; DATA 21100 (or equivalent); DATA 21300 (could be taken concurrently) and DATA 21200 (or equivalent) recommended

DATA 23700. Visualization for Data Science. 100 Units.

This course imparts design knowledge and technical skills around data visualization that data scientists need for analysis and communication. Design principles for data visualization are based on frameworks for thinking about chart construction, empirical research on human perception and cognition, a repertoire of design strategies, and theories about what makes an effective chart. Data Science majors will build practical skills around visualization APIs, computational notebooks, graphics editing software, and technical writing. Topics of interest include visualization software, spatial and visual reasoning, cartography, making data interactive, persuasion and deception, uncertainty communication, and model interpretability.

Instructor(s): A. Kale     Terms Offered: Winter
Prerequisite(s): DATA 11900 or CMSC 14100, DATA 12000 (could be taken concurrently), DATA 21300 (could be taken concurrently)

DATA 24100. Software Engineering for Data Science. 100 Units.

This course is designed to equip students with the practical skills and theoretical knowledge necessary to excel at the intersection of data science and software engineering. Through a hands-on approach, students will delve into the core tools and concepts that form the backbone of this interdisciplinary field, including data modeling, building data pipelines and software development best-practices. Emphasis will be placed on real-world applications, enabling students to work on projects that simulate professional scenarios and challenges. This course is ideal for those looking to deepen their understanding of how data-focused technologies are developed and deployed.

Instructor(s): N. Ross     Terms Offered: Autumn
Prerequisite(s): DATA 12000

DATA 25422. Machine Learning for Computer Systems. 100 Units.

This course will cover topics at the intersection of machine learning and systems, with a focus on applications of machine learning to computer systems. Topics covered will include applications of machine learning models to security, performance analysis, and prediction problems in systems; data preparation, feature selection, and feature extraction; design, development, and evaluation of machine learning models and pipelines; fairness, interpretability, and explainability of machine learning models; and testing and debugging of machine learning models. The topic of machine learning for computer systems is broad. Given the expertise of the instructor, many of the examples this term will focus on applications to computer networking. Yet, many of these principles apply broadly, across computer systems. You can and should think of this course as a practical hands-on introduction to machine learning models and concepts that will allow you to apply these models in practice. We'll focus on examples from networking, but you will walk away from the course with a good understanding of how to apply machine learning models to real-world datasets, how to use machine learning to help computer systems operate better, and the practical challenges with deploying machine learning models in practice."

Instructor(s): Nick Feamster
Prerequisite(s): CMSC 14300 or CMSC 15400
Equivalent Course(s): DATA 35422, CMSC 25422, CMSC 35422

DATA 25900. Ethics, Fairness, Responsibility, and Privacy in Data Science. 100 Units.

This course takes a technical approach to exploring societal issues of ethics, fairness, responsibility, and privacy related to the collection, use, and generalization of data. The course introduces fundamental techniques related to data acquisition, data cleaning, sampling, statistical modeling, experimental design, feature engineering, and modeling with machine learning. It then explores the problems that arise in different ways of performing those tasks, the fairness and bias of machine learning models, data visualizations, and user interfaces. In addition, the course covers anonymization and deanonymization, conceptions of privacy from a number of perspectives (statistical, legal, and philosophical), and compliance with contractual or legal requirements around data. The course concludes by discussing current controversies around the use and misuse of data. Through both programming assignments and discussions, students who complete the course will learn how to design systems that are inclusive and respectful of all data subjects.

Instructor(s): A. Nussbaum, A. Kube     Terms Offered: Autumn Spring
Prerequisite(s): DATA 11900
Note(s): Students may not take DATA 25900 if they have taken CMSC 25900 or CMSC 25910.

DATA 26100. Statistical Pitfalls and Misinterpretation of Data. 100 Units.

This course will provide tools for thinking critically about data and models that constitute evidence, e.g., for the purpose of making predictions, reaching decisions, judging the quality of information, or in the context of scientific research. The course will examine examples of misleading language and graphics and discuss good practices in representing and communicating data. Examples will include such pitfalls as data size effects, false linearity, biased or correlated samples, mistaking correlation for causation, regression to the mean, and Simpson's paradox.

Instructor(s): W. Trimble     Terms Offered: Spring
Prerequisite(s): MATH 13200 and DATA 11800 or CMSC 12100 or CMSC 14100

DATA 27100-27200. Data Science Clinic I-II.

Data Science Clinic I-II

DATA 27420. Introduction to Causality with Machine Learning. 100 Units.

This course is an introduction to causal inference. We'll cover the core ideas of causal inference and what distinguishes it from traditional observational modeling. This includes an introduction to some foundational ideas---structural equation models, causal directed acyclic graphs, and then do calculus. The course has a particular emphasis on the estimation of causal effects using machine learning methods.

Instructor(s): V. Veitch     Terms Offered: TBD
Prerequisite(s): [STAT 24500 or STAT 24510 or STAT 27725] with a grade of B or higher or consent of instructor.
Equivalent Course(s): STAT 27420

DATA 27751. Trustworthy Machine Learning. 100 Units.

Machine learning systems are routinely used in safety critical situations in the real world. However, they often dramatically fail! This course covers foundational and practical concerns in building machine learning systems that can be trusted. Topics include foundational issues---when do systems generalize, and why, essential results in fairness and domain shifts, and evaluations beyond standard test/train splits. This is an intermediate level course in machine learning; students should have at least one previous course in machine learning.

Terms Offered: TBD
Prerequisite(s): STAT 27700 or STAT 37710 or consent of instructor.
Equivalent Course(s): STAT 27751, STAT 37787