INFX 598 C/D: Introduction to Data Science

Winter 2013
University of Washington School of Information
Lectures: Tuesday and Thursday 1:30-3:20, MGH 420

Course Description | Prerequisites | Course Outline | Grading | Assignments | Calendar

Instructor: Prof. Joshua Blumenstock, Suite 370 Mary Gates Hall

This course offers students an introduction to the growing field of "Data Science" as practiced by leading data scientists in industry and research. As “big data” become the norm in modern business and research environments, there is a growing demand for individuals who are able to make decisions and derive meaningful insight from large-scale, heterogeneous data. This requires a diverse mix of skills, from data munging and ETL; to distributed storage and computing; to machine learning and econometrics; to effective visualization and communication.

Through a combination of hands-on exercises and guest lectures by experts in the field, this course provides an overview of several key concepts, skills, and technologies used by practicing data scientists. While the lectures will be intelligible to a general audience, successful completion of assignments requires college-level exposure to statistics and programming.

A data scientist is often referred to as someone who knows more statistics than a computer scientist and more computer science than a statistician. Students enrolled in the course must have college-level exposure to both statistics and programming. Students who do not meet the following requirements, and in particular the programming requirement, will find it difficult to satisfactorily complete problem sets, and should consider taking the course at a later date.

Programming: Students should be able to comfortably program in a high level programming language like Java, python, php, or C/C#/C++. Note that html, javascript, and VBA are not sufficient in this context. “Comfortably” implies that students should be able to write simple programs from scratch, like a web scraper, or a text parser, or a simple game of scrabble or tic-tac-toe. Real-world programming experience (e.g. an intensive summer programming internship) will be as useful than formal CS coursework.

Statistics: Students should have had prior exposure to regression methods and should understand concepts of hypothesis testing and statistical significance. Experience with R or MatLab will be especially valuable in completing the problem sets.

Students are required to read the material assigned each week before section, and will be expected to actively engage in discussion of these materials. Course grades will based on a final group project, two problem sets, discussion leadership, and overall classroom participation. The two problem sets will be programming assignments, using R and Python. The final project will be done in groups of 3-4 students; further details will be provided at the start of the quarter.

• Group final project: 40%
• Technical problem set 1: 20%
• Technical problem set 2: 20%
• Class Discussion Leadership: 10%
• Class Participation: 10%
• Extra Credit: 5%

Grading Policy

• All assignments are to be submitted on Canvas by 12pm on the due date.
• Assignments turned in *up to* 24 hours after the due date will be penalized 20%.
• Any assignments turned in more than 24 hours late will receive no credit.

A Note on Programming Languages

Most in-class code demonstration will involve either R or Python. Guest lecturers may give examples in other languages but will explain what the code means. Homework assignments will generally require R or Python. If you would prefer to use a different language successfully, you may do so, but you will be “on your own” and should not expect technical support.

Academic Integrity Policy

Discussion with instructors and classmates is encouraged, but each student must turn in individual, original work and cite appropriate sources where appropriate.

Readings and Class Discussion

Required and optional readings will be announced in class and posted on the course website. Students will be assigned by the instructor into groups of 2-4, and each group will be responsible for guiding classroom discussion for a given week (two classes). Unless otherwise noted by the instructor, a discussion will last roughly 10-20 minutes, and will consist of (i) a 5-10 minute summary of all required and optional readings for that week, and (ii) 10-15 minutes of open discussion, moderated by the discussion leaders. Discussion leaders are also responsible for posting a 500-1000 word entry on the class blog/wiki that summarizes the required and optional readings for the week, as well as any salient points from the in-class discussion. This post must be made by 12pm on the Friday following second class.

Over the course of the quarter, we will host several prominent visiting speakers. On days when one or more speakers are present, there will be no discussion of the readings (in such instances, all readings for a week will be discussed on the day without a speaker). Instead, the discussion leaders will be responsible for researching the visiting speaker, meeting the speaker and walking him/her to the classroom, and moderating the classroom discussion and Q&A following the speaker’s presentation. Moderators should prepare several intelligent, challenging questions to ask the speaker immediately after the talk finishes, and if ever there is a lull in the discussion.

Group Project (40%): Students will form groups of 2-4, and complete one of the following assignments. Each group must submit a 200-300 word description of the proposed project by 12pm on January 24.

• Analysis of a dataset “in the wild”
• Sectoral Analysis

Problem Set 1 (20 points): Programming Assignment 1: Data processing and basic analytics
The goal of this problem set is to familiarize yourself with implementing basic statistical analysis on a small dataset using R. It is important that you read Chapters 1-3 carefully, and work to understand exactly how the authors are using R to manipulate the underlying data. If you merely hack your way through the exercises without working through the material in the chapters, you will come away with a very incomplete understanding of how to perform such analysis.
Assignment: After reading Chapters 1-3 (pp 1-64) of Everitt & Hothorn, complete the following exercises (2 points each, 18 points total) : 1.1, 1.2, 1.3, 1.4; 2.1, 2.3, 2.4; 3.1, 3.2. Extra credit (2 points): 3.3. Submit two files:

• A polished .pdf with your solutions. 2 points will be given for overall presentation, so take care to appropriately label your graphs and tables. At the top of your submission, write your name and the names of any students with whom you worked, if applicable. Also write an estimate of how long, in hours, it took you to complete the exercise -- this will not affect your grade, it is for calibration.
• Your commented R code as a plain text file

Problem Set 2 (20 points): Programming Assignment 2: Advanced analytics and visualization

• Details TBD

Classroom Discussion Leadership (10%):

• See description in above section, “Readings and Class Discussion”

Extra Credit (5%): Several popular books have been written on topics thematically relevant to this course. Excerpts from some of these books are required reading. For extra credit, write a short (400-800 words) review of one of the books listed below. The review should summarize the core points of the book and provide a critical analysis that draws attention to strengths and shortcomings of the book. Please do not review a book that you have previously read.

• Davenport & Harris. Competing on Analytics
• Davenport, Harris, & Morison. Analytics at Work
• Ian Ayres. Super Crunchers: Why Thinking-by-Numbers Is the New Way to Be Smart
• Steven Baker. The Numerati
• Thomas Redman. Data Driven: Profiting from Your Most Important Business Asset
• Sasha Issenberg: The Victory Lab
• Nate Silver: The Signal and the Noise