Introduction to Data Science Using R

Version: April 4, 2018

Prof. Darin Christensen1 1

6341 Public Policy
[email protected]
(310) 825-7196
1 Course Information

• Class Location & Schedule: Public Affairs 2250; TR, 5-6:15p

• Section Location & Schedule: Public Affairs 4357; W, 3-3:50p

• Office Hours: W, 8:00-10:30a; F, by appointment.

• Course Website: https://piazza.com/ucla/spring2018/pp290dis3/

home

• Prerequisites: None.

2 Course Overview

There is a growing demand that public policy be “data-driven.” The

White House, to take a prominent example, appointed its first-ever
Chief Data Scientist in 2015. While ten weeks isn’t enough time to
prepare you for that job, this class introduces you to a set of common
steps that data scientists use to transform a spreadsheet into a useful
description or illustration. We cover acquiring, cleaning, merging,
managing, summarizing, and visualizing quantitative data.2 While 2
If you’re interested in effective data
these tasks can be accomplished using different software, this course visualization consider taking Prof.
Reber’s Arguing with Data course.
demonstrates how to use R, a free and popular statistical program.

3 Learning Goals

There are several goals for this course. By the end of the quarter, you
should be able to:

1. Write down a recipe (i.e., a set of ordered steps) for transforming a

dataset into a table or figure that describes a pattern;

2. Identify and apply functions/packages in R for accomplishing the 3

You won’t be expected to memorize
every R function, but rather identify
steps of this recipe;3
sources for useful code snippets and
understand how to adapt those snip-
3. Combine data from multiple (unrelated) sources; and pets.

4. Apply best practices in writing code, including using clear naming

conventions, commenting steps, and avoiding repetition. These
conventions help ensure that code is error-free and legible to our
colleagues and future selves.
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 2

An important disclaimer: like any new language, R is challeng-

ing to learn, and you’ll spend time fighting with the program. In
those darker moments, try to remember two things. First, you are
not alone; your peers and I can both relate and also offer help when
you’re hitting a wall.4 Second, remember that the first goal of this 4
Our Piazza message board is a great
course has nothing to do with R. Writing down the recipe that trans- place to post coding questions and get
quick responses.
forms data into insights requires careful, ordered reasoning (i.e.,
“algorithmic thinking”), not a specific piece of software.

4 Assignments

• Participation (10%): There are several ways to participate in and

outside of class. I encourage you to raise questions and make
comments during our class meetings. In addition, I will reward
students for assisting peers during in-class group work, as well as
posting questions or responses to our Piazza discussion board.

• Weekly Problem Sets (30%): These assignments serve two pur-

poses. First, they allow you to apply the reasoning and coding
skills we cover in class. Second, some problems will help you get
started on your final project. While there will not be much read-
ing, you will need to set aside time for problem sets.
All problem sets must be completed in R Markdown (discussed 5
I will post detailed solutions online
below). These assignments will be evaluated according to the but cannot grade every problem. Please
come to office hours or schedule an
following scale:5
appointment if you have any questions.

– X : complete
– X− : submitted, but incomplete
– 0 : not submitted or less than half complete

• Midterm Exam (35%): The midterm exam offers an opportunity

to assess your progress on the first two learning goals. As such,
part of the exam will test your problem solving skills; another
component will ask you to write R code.
The mid-term will be conducted during class time. However, you can take
the exam from the location of your choice. You will need R to complete the
exam.

• Final Project / Hackathon (25%): As a group of up to four stu-

dents, you will write a quantitative report based on one of the
datasets listed below. As with other “hackathons,” this is a friendly
competition between groups to squeeze insights out of the same
data.
This project will allow you to practice the real-world tasks of ma-
nipulating, and analyzing quantitative data and clearly conveying
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 3

your approach and results. In addition to the written report, you

will (time permitting) briefly present your findings during the last
week of class. More detail about the format and rubric for the final
project will be provided.

5 Course Policies

5.1 Late Submissions

I don’t want you to fall behind, so I will not accept late homework
assignments. However, I understand that there are forces beyond
your control, so I will drop your lowest homework grade.
Final projects will be docked a letter grade for every day that they
are late. For example, an A project submitted in the first 24 hours
after the deadline will receive an B. If that same project is submitted
24 to 48 hours after the deadline, the grade drops to a C.

5.2 Academic Accommodations

Academic Accommodations. Students needing academic accommo-
dations based on a disability should contact the Center for Accessible
Education (CAE). When possible, students should contact the CAE
within the first two weeks of the term as reasonable notice is needed
to coordinate accommodations. For more information visit.

5.3 Academic Integrity

Please review UCLA’s rules related to academic integrity. If you’re
feeling overwhelmed or are unsure about the collaboration policy,
please speak with me; don’t risk violating the honor code. Unfortu-
nately, past MPP students have been investigated and punished by
the Office of the Dean of Students.
Collaboration Policy:
• Problem Sets: You can work with peers (in groups of four or less).
However, you must write every line of code and text that appears
in any assignment that you submit. In short, never copy code or
use other students’ words to describe your approach or results.

• Midterm Exam: This is an individual assessment that helps us

identify any knowledge gaps. I want to know what you know, so
there’s no collaboration.

• Final Project: You will work with a partner on your final project.
You are responsible for ensuring that no written work is copied or
summarized from other sources without appropriate attribution.
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 4

6 Resources

6.1 Mental Health

Let Public Policy’s Student Affairs Officer (SAO) Annie Kim know if
you are having academic or personal problems that affect your ability
to participate fully in your education. If you feel comfortable, let me
know as well.
UCLA Counseling and Psychological Services offers services and
programs in a confidential environment to promote mental health
and wellness.

6.2 Immigration
The Bruin Resource Center’s (BRC) Undocumented Student Program
offers caring and personalized support to undergraduate and gradu-
ate undocumented students.
Even if you are not undocumented, you may be able to get legal
help for a family member. The USP office provides immigration legal
services to students and their family members through a partnership
with the UC Undocumented Legal Services Center.

6.3 Software
You will need to acquire some software before the first day of class:
• You will need to download and install R from this website: https:
//cran.r-project.org/.

• Use RStudio, which can be downloaded here: https://www.

rstudio.com/products/RStudio/. RStudio provides a nice in-
terface for writing and executing R code and composing R Mark-
down documents.

• You will use R Markdown (http://rmarkdown.rstudio.com/) to

write your homework assignments and final project. R Markdown
6
I won’t devote class time to teaching
allows you to create well-formatted documents that embed R code Markdown, but I can answer ques-
and outputs.6 tions and point you to useful online
resources.
• To compile R Markdown to PDF, you’ll need to install MacTeX
(Mac) or MiKTeX (Windows).

6.4 Using R
The resource we will use most is R for Data Science by Wickham
and Grolemund. An online version of the text is available here
(http://r4ds.had.co.nz/) or you can purchase the book once it’s
published (http://goo.gl/EUUNeh).
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 5

There are many online and print resources for learning R. I’ve
included a few below, but a Google search will turn up many others.

Online:

• Interactive Tutorials: Try R

• Examples: UCLA’s Institute for Digital Research and Education

• Examples: Cookbook for R

• Crowd-sourced Solutions: Stack Overflow

Print:

• R for Dummies (http://goo.gl/mel0fo)

• R Cookbook (http://goo.gl/Hh1REr)

We are going to rely heavily on two R packages: dplyr for data ma-
nipulation (https://goo.gl/Nc5gIV), and ggplot2 for visualization
(http://docs.ggplot2.org/current/).

6.5 Data Sources for Final Project / Hackathon

I recognize that there are diverse policy interests in the course. In an
effort to accommodate this, I’ll permit groups to choose one of two
datasets for their final project:

1. World Bank Geocoded Research Release: “This geocoded dataset

release represents all World Bank projects in the IBRD and IDA
lending lines approved from 1995-2014. This dataset tracks 5684
geocoded projects across 61243 locations.”

2. Database on Ideology, Money in Politics, and Elections: “The

resulting database contains over 130 million political contributions
made by individuals and organizations to local, state, and federal
elections spanning a period from 1979 to 2014.”
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 6

7 Course Schedule

Legend: ?: Required; • Suggested.

Caveat: This schedule is approximate, and we may drop or reorder sections as the quarter proceeds.

0 Before the First Class

? Complete the Try R interactive tutorials; and

? Download, install, and open R, RStudio, and MiKTeX (Windows) or MacTeX (Mac).
? Open and compile this R Markdown file. If this compiles to a PDF, then you’re ready for class.
? Sign up for the course website at piazza.com/ucla/spring2018/pp290dis3.

1 Introductions (4/3)

? Chapters 1 and 2: Introduction. Garrett Grolemund and Hadley Wickham. R for Data Science. O’Reilly
• Mike Loukides. What is data science?, June 2010. URL https://www.oreilly.com/ideas/what-is-data-science
• Sharp Sight Labs. Why you should learn R first for data science, January 2015. URL http://www.
r-bloggers.com/why-you-should-learn-r-first-for-data-science/
• Paul Curzon. A recipe for programming, 2014. URL https://teachinglondoncomputing.files.
wordpress.com/2014/07/cs4fnissue16_pr_6_1.pdf

2 Description

2.1 Visualizing Data (4/5, 4/10)

? Chapter 3: Data Visualisation. Grolemund and Wickham.
? Jonathan A Schwabish. An Economist’s Guide to Visualizing Data. The Journal of Economic Perspectives,
28(1):209–34, February 2014
• H Wickham. A layered grammar of graphics. Journal of Computational and Graphical Statistics, 19(1), 2010

2.2 Manipulating Data Frames (4/12, 4/17)

? Chapter 5: Data Transformation. Grolemund and Wickham.

3 Wrangling

3.1 Importing Data (4/19)

? Chapter 11: Data Import. Grolemund and Wickham.

3.2 Cleaning (or “Tidying”) Data (4/24)

? Chapter 12: Tidy Data. Grolemund and Wickham.
 I N T R O D U C T I O N TO DATA S C I E N C E U S I N G R 7

3.3 Working with Dates and Strings (4/26, 5/8)

? Chapter 16: Dates and Times. Grolemund and Wickham.
? Chapter 14: Strings. Grolemund and Wickham.
[Midterm Exam (5/3)]

3.4 Merging, Joining, and Appending (5/10)

? Chapter 13: Relational Data. Grolemund and Wickham.

4 Coding

4.1 Functions (5/15)

? Chapter 19: Functions. Grolemund and Wickham.

4.2 Iteration (5/17, 5/22)

? Chapter 21: Iteration. Grolemund and Wickham.
? Chapter 20: Vectors. Grolemund and Wickham.

5 Special Topics (Time Permitting)

5.1 GIS in R (5/24)

• Using the sf Package. Here’s a vignette that demonstrates how to use the package.

5.2 Basic Models (5/29)

? Chapter 22-23: Introduction & Model Basics. Grolemund and Wickham.

6 Skill Consolidation (5/31, 6/5)

7 Final Presentations

[Final Project Due (11 June 2018 by 11:59 p.m.)]