NYU Computational Social Science

Certificate Program

Design

Sargent designed the certificate program in close collaboration with Chase Coleman and Spencer Lyon, lead developers at QuantEcon, an open source website devoted to teaching how to put Python and Julia at the service of the social sciences. At NYU, Coleman, Lyon, and Sargent will team teach all four courses in the sequence, as well as the online summer “pre-class”.

Purposes

Advancements in computational power over the last 30 years have led to significant steps forward in how we approach social science, giving rise to a new class of social scientists: computational social scientists. Computational social scientists work throughout the public and private sector: in universities, private research institutions, and business enterprises such as tech companies like Amazon, Facebook, Google, and Uber.

An NYU Computational Social Science certificate program aims to prepare students for either a graduate program in the social sciences or for a career as a data analyst or computational social scientist.

The certificate program consists of four one-semester long classes together with an online preparatory class to be completed during the summer preceding the certificate program sequence. Students with qualifying experience or coursework can opt out of the “pre-class”.

Prerequisites and location

No prior knowledge of computer programming or machine learning is assumed. The Online Foundations Course (see below) will provide the basic computer programming skills required to succeed in subsequent courses in the sequence. The program does assume familiarity with elements of multivariate calculus, linear algebra, statistics, and probability.

The sequence will mix online and in-person classes with ample student-teacher interactions accomplished either in person or online.

Course Descriptions

Before program (online)

ECON GA 4001, 0 credits
Online Foundations Course: The foundational course of the sequence is an online, self-paced course typically completed over 12 weeks. This course covers skills that will be used for the remainder of the year and we assume students are familiar with this material in the fall. The class introduces students to the Python programming language and elementary software engineering tools that will be applied in subsequent courses in the sequence.(Syllabus)

Semester 1 (Fall)

ECON GA 4002, 3 credits
Mathematical Foundations for Computational Social Science: This course reviews essential mathematical tools — such as calculus and linear algebra — and introduces students to foundational concepts of random variables, model building, and model estimation (both frequentist and Bayesian). The goal of this course is two-fold: (1) to empower students to build models to discuss the world around them, and, (2) to foster mathematical maturity so that students can teach themselves mathematically useful material as needed. (Syllabus)

ECON GA 4003, 3 credits
3 credits Data Tools for Computational Social Science: This course arms students with cutting-edge data manipulation and management tools. The class relies heavily on the Python pandas package and emphasizes that “real-world” data are messy. The course provides students with tools for analysis in such environments. (Syllabus)

Semester 2 (Spring)

ECON GA 4004, 3 credits
Dynamic Models for Computational Social Science: This course covers dynamic models and their uses for guiding decisions. Topics include dynamic programming, time-series analysis (both Bayesian and frequentist), Markov models, and Hidden Markov Models. Many of these topics are the focus of cutting-edge research. Examples include text analysis. This course empowers students to apply these tools in academic, government, and industry research. (Syllabus)

ECON GA 4005, 3 credits
Machine Learning for Computational Social Science: When applied correctly, “machine learning” tools allow individuals to approximate complicated outcomes in the real world. However, when applied carelessly, these tools generate misleading findings. This course covers supervised learning (both regression and classification) , reinforcement learning, and model selection via validation procedures. The course prepares students to apply classical and cutting-edge machine learning techniques to problems in the social sciences. The course presents a principled approach that adheres to best practices and encourages understanding and transparency. (Syllabus)

Course Sequence

Semester Course 1 Course 2
Summer Online Foundations Course
Fall Mathematical Foundations for Computational Social Science Data Tools for Computational Social Science
Spring Dynamic Models for Computational Social Science Machine Learning for Computational Social Science