Introduction to Human-Centered Data Science 

In this course, students will be introduced to the technical, social, and ethical considerations in the field of data science, including data security, governance, and privacy. Particular focus will be placed on data scientists’ responsibility to create effective and inclusive solutions that are responsive to the needs, values, and perspectives of people.

This course will also introduce the themes and skills that will be developed in subsequent courses. Specifically, students will: learn about trends and advances in data science (e.g., A.I. and cloud computing); be exposed to data cleaning procedures for various types of commonly used data; learn about project lifecycle planning and execution; and will learn about the steps in a typical data science project (e.g., question framing, data collection, cleaning, exploration, modeling, interpretation of findings). Moreover, instruction will focus on developing students’ skills relating to project management, problem framing, communication, and project execution.

Applied Mathematics for Data Science

This course will provide students with a strong foundation in mathematical tools relevant to data science. Selected topics from calculus, vector spaces, matrix algebra, numerical optimization, and probability theory will be covered. These tools will help students understand and solve data science problems and work with emerging methods and techniques in this rapidly growing field. A refresher in calculus will provide a foundation in mathematics necessary to understand data science concepts. Topics will include matrix algebra and vector spaces, which are essential for understanding mathematical models and statistical methods used in data science, and numerical mathematics and optimization, which are integral for understanding model training and efficiency. Additionally, skills in this area can help detect overfitting by providing a way to assess the quality of a model’s fit to data. Finally, instructional units on basic probability theory will lay the foundation for the subsequent classes on statistical models for data science.

Programming for Data Science

This class will introduce essential programming concepts, data structures, and techniques focusing on data science applications. Students will learn what it means to write high-quality code. Additionally, topics such as testing, debugging, and exception handling will be taught. By the end of the course, students will have reviewed Python programming and be able to apply their knowledge to various data science problems. Topics include functions, recursion, loops, list comprehensions, elementary data structures, and reading and writing files.

General Linear Models 

This course addresses the construction, interpretation, and application of general linear statistical models. Students will learn data and sampling distributions, A/B testing, significance testing, simple and multiple regression, power analysis, parameter estimation, partial associations, regression diagnostics, missing data analysis, remedial analysis, model misspecification, exploratory versus confirmatory model building, categorical data coding, dummy-variable regression, mediation, moderation, polynomial models, prediction, cross-validation, and regularization including ridge regression and Lasso regression. 

By applying methods learned in the class to real data, students will learn to be responsible and careful data scientists. Students will learn to be conscientious about data sources, paying particular attention to individuals and populations that may be systematically excluded.

Database Systems and Data Preparation

This course will provide a foundation in discrete mathematics, data structures, algorithmic design and implementation, and computational complexity. Students will be introduced to relational databases focusing on learning SQL with the Postgres database. Topics include schemas, indexes, query efficiency, server-specific navigation functions, and queries with grouping, ordering, sorting, collapsing, and joins. Distributed data storage and processing techniques for parallel processing (MapReduce) and their implementation (Hadoop) will be covered, as well as strategies for accessing unstructured data and handling streaming data.

Multivariate Statistical Analysis

Modern statistics and data science often deal with high-dimensional data, and multivariate methods are used to handle these types of data. Approaches to supervised and unsupervised learning with multivariate data will be discussed. This course will cover exploratory data analysis, binary logistic regression, multinomial logistic regression, ordinal logistic regression, discriminant analysis, support vector machine (SVM), K-Nearest Neighbors (KNN), Multivariate Analysis of Variance (MANOVA), principal component analysis, exploratory factor analysis, structural equation modeling, cluster analysis. For each topic, theory and applied analysis will be covered, and students will use the statistical packages R and SPSS to calculate and report multivariate statistics. 

For many issues in public life, statistics and its application is a critical scientific approach to understanding how actions and policies affect people’s lives. Good policies advance social justice, and policy debates should be informed by responsible practitioners, reliable data sources, and statistical data analysis results. In this course, students will learn to use reasonable and appropriate methods to handle data analyses and to pay attention to data sources, outliers, and quality analytics.

Data Visualization and Storytelling

This course will introduce the principles for guiding the development of compelling stories based on data and how to use visualizations to assist in articulating key findings. Visualization and storytelling are important tasks in the lifecycle of a data science project and essential skills to communicate effectively. Data visualization allows data scientists to take complex data and present it in a way that is easy to understand, while storytelling allows them to add context to their findings and explain their importance. Students will learn various data visualization techniques for reporting purposes and the widely used software tools for implementing these visualizations. Students will also learn how to create interactive reports based on dashboards deployed on cloud servers to maximize client engagement. 

Applied AI and Machine Learning 

This class will provide a broad overview of various approaches to machine learning, including supervised and unsupervised learning. Students will learn about the fundamental algorithms used to train computers to learn. The course will also expose students to different application areas where data-driven decision-making is aided by machine learning (e.g., text classification, image recognition, and predictive modeling). Students will use the Python programming language and machine learning libraries (e.g., scikit-learn) to solve authentic problems. While working with authentic datasets, students will also learn about bias, accountability, and trust issues that arise when conducting human-centered data science.

Deep Learning and AI 

This course will introduce deep neural networks. Students will learn foundational concepts such as stochastic gradient descent and backpropagation and how they are used to train network models. Students will also gain practical experience using Python libraries (e.g., TensorFlow, Keras) to implement various deep learning architectures. Different types of artificial neural network architectures will be covered so that students will learn about classes of applications with an eye toward human-centered data science in this class. These include convolutional neural networks, recurrent neural networks, and long short-term memory networks. Students will learn how to choose different architectures depending on the tasks being addressed.

Applied Data Science Capstone I & II

In Capstones 1 and 2, students will complete a significant data science research project that requires executing learned skills in problem framing, problem-solving, and visualization. While working with a faculty member and internal and/or external partners, students will be responsible for identifying questions to investigate, data sources, and applying analytical techniques. Students will prepare a written report of their results and deliver a presentation to faculty and partners. The project synthesizes the statistical and computational challenges of solving complex real-world problems with social issues. The course activities focus on a two-semester-length data science project supervised by a faculty member and in collaboration with Boston College and/or external partners. At the conclusion of these capstone projects, students will have accumulated hands-on experience working with clients, applying methods, and developing materials and presentations that can be shared with others.

Natural Language Processing

This class on natural language processing (NLP) will cover issues such as topic modeling, text summarization and classification, sentiment analysis, large language models, and automatic scoring of a student’s written response, including essay scoring. NLP provides essential methods for dealing with large amounts of text data created when sourcing information from the web and other large text corpora. Practical exercises in this class will help improve student awareness around fairness, bias, and human-centered data science by providing students with opportunities to learn and apply methods for dealing with large amounts of real-world text data. The techniques taught in this class will help to improve the accuracy and fairness of predictions made by machine learning models and can also help to improve the interpretability of those models.