COURSE NUMBER: MBA217.2
COURSE TITLE: Big Data and Better
Decisions
UNITS OF CREDIT: 3
INSTRUCTOR: Jonathan Kolstad, Paul Gertler
E-MAIL ADDRESS:
jkolstad@berkeley.edu
CLASS WEB PAGE LOCATION:
bcourses.berkeley.edu
PREREQUISITE(S): Data and
Decisions
CLASS FORMAT: Lectures with in
class cases and work sessions on problem sets.
REQUIRED READINGS: Mixture of text
books (most available online), papers and cases.
BASIS FOR FINAL GRADE: Two exams,
class participation
ABSTRACT OF COURSE'S CONTENT AND OBJECTIVES: This course will introduce students to advanced methods for data driven decision making in business. Building on the content in Data and Decisions, this course will cover methods designed to provide evidence for two types of fundamental business issues. The first is forecasting and the second is evaluating alternative possible strategies. The course is intended to train business leaders to i) understand the value of data-based decision making ii) evaluate analytics tools and products and iii) conduct richer analysis of randomized and naturally occurring experiments. Topics include designing randomized controlled trials in the field, evaluating natural experiments (e.g. differences-in-differences, instrumental variables and regression discontinuity) and machine learning tools for forecasting (e.g. linear regularization, tree models and random forest). The course work will include problem sets evaluating experimental and non-experimental data, writing code and making business decisions using the results. The goal of the course is not to train you as a Data Scientist but to be able to read and evaluate empirical/analytic approaches and products from any level including reading code to evaluating experimental design. All problem sets will be in R and the course will introduce the programming language.