Foundations of Data Science with R
This foundational course offers a full-spectrum introduction to data science and data science workflows, emphasizing data as a source of value creation in the enterprise. The R programming environment serves as the implementation vehicle in support of essential data science activities – data exploration and visualization, data wrangling, predictive modeling, model deployment, and communication. The R programming environment, along with Python, is among the most important tools in the data scientist’s toolbox and this course will feature tools and a style of programming inspired by the popular tidyverse ecosystem – ggplot2 for data visualization, dplyr and tidyr for data wrangling. Students will master elements of the data science workflow through a series of short R programming exercises reinforced by a full-spectrum, integrative final project. Presentation skills are an ever-present theme as students are challenged, through every stage of analysis, to communicate managerial relevance and value to the enterprise.
Research Design, Visualization, and Presentation
It’s one thing to conduct an analysis, it’s another to convince someone to change their behavior based on this analysis. In this course, students will study theories of visualization, communication and presentation with the purpose of translating technical results into actionable insight. Using a mix of case studies and code, the course begins with an examination of how to ask good research questions. It then covers the psychology of communication and the construction of compelling visualizations. Finally, students are tasked with writing and presenting their work in a manner suited to a non-technical audience.
Fundamentals of Data Engineering
Data management is core to both applied computer science and data science. This includes storing, managing, and processing datasets of varying sizes and types. This course introduces students to the various ways in which data is stored and processed including relational databases, file-based databases, cloud-based storage and data streaming. A key component of the course is learning which architectures fit which types of data science problem (and the strengths and weaknesses of each). Students will learn to work with data that is both clean and structured, and dirty and unstructured.
Applied Machine Learning
Machine learning is becoming a core component of many modern organizational processes. It is a growing field at the intersection of computer science and statistics focused on finding patterns in data. Prominent applications include personalized recommendations, image processing and speech recognition. This course will focus on the application of existing machine learning libraries to practical problems faced by organizations. Through lectures, cases and programming projects, students will learn how to use machine learning to solve real world problems, run evaluations and interpret their results.
Time Series and Longitudinal Data Science
Data over time present unique challenges and opportunities for the data scientist. Just the specification of time is complicated owing to time zones and daylight savings time and models of data over time must confront trends and patterns that need be understood and exploited. The first part of the course examines forecasting and time series as data science problems. We continue to non-quantitative data over time and combining data over time across multiple, potentially quite different, units. Throughout, we will emphasize forecasting and the evaluation of forecasts with a balance of quantitative and visual tools.
Practical Applications of Python for Data Science
A relatively fast-paced introduction to the Python programming language and its use in data science. Topics include typical Python objects and control structures with a special focus on data ingestion and manipulation. Special attention will be devoted to Python’s system of packages for data analysis including Pandas, SciKit, Numpy and methods to move data between Python and common statistics software like R and Stata. Weekly programming exercises complement in-class lessons, and a final project gives students the opportunity to incorporate the various skills they’ve been learning into a reproducible report analogous to those used in industry to present insights from a data science effort.
Data Ethics, Policy and Human Beings
This course explores the legal, policy, and ethical implications of data. These types of issue arise at each stage of the data science workflow including data collection, storage, processing, analysis and use. Armed with legal and ethical guidelines, students are then confronted with topics including privacy, surveillance, security, classification, discrimination, decisional-autonomy, and duties to warn or act. Using case studies and a lecture-discussion format, the course will address real-world problems in areas like criminal justice, national security, health, marketing and politics.
Over the course of the semester, students will propose, plan and execute an actual data science project. Run as an independent study during the student’s last term, the project provides an opportunity to integrate all of the core skills learned throughout the program, and to develop a portfolio piece that can help with students’ career aspirations. Projects must be consequential in nature—i.e., have a real (or potential) impact on some organization, or the world. Grades will be based on assessments by both the faculty advisor and those (potentially) impacted by the project’s results. Data sets must be selected by the student either from a public repository or from the company for which they work and approved by the course instructor within the first two weeks of the term.