POLS 1600

Data and Measurement

Updated May 31, 2024

Overview

Class Plan

  • Logistics (15 minutes)
    • Announcements
    • Feedback
  • Class plan (60 minutes)
    • Introduction to R, R Studio and Quarto
    • Loading and Looking at Data in R
    • Transforming, Recoding, and Cleaning Data in R
    • Describing Data in R
    • Exploring Covid-19 Data for Lab

Annoucements

  • If it’s your first time here you’ll need to work through Software Setup to follow along today

    • Talk to me after class if you’re having installation issues

  • If you’re still on the waitlist on CAB, speak to me after class

Announcements

Tutorials

Once you’ve done the following

remotes::install_github("rstudio/learnr")
remotes::install_github("rstudio-education/gradethis")
remotes::install_github("PaulTestaBrown/qsslearnr")

You can see the available problem sets by running the following code in your console:

learnr::run_tutorial(package = "qsslearnr")

Tutorials

And start a specific tutorial by running:

learnr::run_tutorial("00-intro", package = "qsslearnr")

Please upload tutorials 00-intro and 01-measurement1 to Canvas by Friday

You’ve committed a murder

Why

Why do you ask?

It’s cousin Nick’s fault…

  • Funny icebreaker, but lots of assumptions…

    • You’re not a murderer

    • You don’t know someone who’s committed a murder or been murdered

    • You’ve got a mom and dad

Why do you ask?

  • How might we make this question better?

    • Use a screener question
      • “Would you feel comfortable…”
    • “Pipe” in responses from a prior question
      • “Who are two people who raised you…”

What questions we ask and how we ask them matters

Hopes and Dreams, Fears and Worries

What are we excited about?

  • Engaging with social science
  • Learning statistics and math
  • Learning to code

What are weworried about?

  • Engaging with social science
  • Learning statistics and math
  • Learning to code

Introduction to R, R Studio and Quarto

Overview

  • R, R Studio and Quarto

  • Getting set up to work in R

  • Basic Programming in R

R, R Studio and Quarto

  • R is an open source statistical programming language (cheatsheet)

  • R Studio is an integrated development environment (IDE) that makes working in R much easier (cheatsheet)

  • Quarto is a publishing system that allows us to write and present code in different formats (cheatsheet)

General Tuesday Workflow

  • Go to https://pols1600.paultesta.org

  • Go to class content for current week

  • Open slides in browser

  • Open R Studio

  • Create .qmd file titled wk01-notes.qmd and save in course folder

  • Get set up to work

  • Take notes and follow along

Let’s create a .qmd file

Three components of a .qmd

  • Control output with YAML header
  ---
  title: "Title here"
  author: "Your name"
  format:
    html:
      toc: true
  ---
  • Write code Blocks/Chunks
```{r}
#| echo: true
2+2
```

  • Describe code using Markdown

    • See Help > Markdown quick reference

The Basics of R

  • R is an interpreter (>)

  • “Everything that exists in R is an object”

  • “Everything that happens in R is the result of a function”

  • Data come in different types, shapes, and sizes

  • Packages make R great

R is an interpreter (>)

Enter commands line-by-line in the console

  • The > means R is a ready for a command

  • The + means your last command isn’t complete

    • If you get stuck with a + use your escape key!

  • Send code from .qmd file to the console:

    • cntrl + Enter (PC) | cmd + Return (Mac) -> run current line
    • cntrl + shift + Enter (PC) | cmd + shift + Return (Mac) -> run all code in current chunk

R is a Calculator

Operator Description Usage
+ addition x + y
- subtraction x - y
* multiplication x * y
/ division x / y
^ raised to the power of x ^ y
abs absolute value abs(x)
%/% integer division x %/% y
%% remainder after division x %% y

R is logical

Operator Description Usage
& and x & y
| or x | y
xor exactly x or y xor(x, y)
! not !x

R is logical

x <- T; y <- F

x == T
[1] TRUE
x == T & y == T
[1] FALSE
x == T | y == T
[1] TRUE
!x
[1] FALSE

R can make comparisons

Operator Description Usage
< less than x < y
<= less than or equal to x <= y
> greater than x > y
>= greater than or equal to x >= y
== exactly equal to x == y
!= not equal to x != y
%in% group membership* x %in% y
is.na is missing is.na(x)
!is.na is not missing !is.na(x)

Everything that exists in R is an object

  • The number 5 is an object in R
5
[1] 5
  • We can assign the object 5, the name x, using the assignment operator <-+
x <- 5 # Read this as "x gets 5"
  • Now if we tell R to show us x, we’ll get
x
[1] 5
print(x)
[1] 5

Data come in different types

Data come in different types

# Create some data

# Numeric
x <- 2 # Double
y <- 6L # Integer

# Logical
only_two_types_of_people <- TRUE 

# Character
me <- "Paul"

# Factor
grades = factor(c("A","B","C"))
# What type are they?
class(x)
[1] "numeric"
class(y)
[1] "integer"
class(only_two_types_of_people)
[1] "logical"
class(me)
[1] "character"
class(grades)
[1] "factor"

Data come in different “shapes” and “sizes”

Source: Gaurav Tiwari

Name “Size” Type of Data R code
scalar 1 numeric, character, factor, logical x <- 5
vector N elements: length(x) all the same v <- c(1, 2, T, "false")
matrix N rows by columns K: dim(x) all the same m <- matrix(y,2,2)
array N row by K column by J dimensions: dim(x) all the same a <- array(m,c(2,2,3))
data frames N row by K column matrix can be different d <-data.frame(x=x, y=y)
tibbles N row by K column matrix can be different d <-tibble(x=x, y=y)
lists can vary can be different l <-list(x,y,m,a,d)

Everything that happens in R is the result of a function

  • You’ve already seen and used some R functions
    • the <- is the assignement operator that assigns a value to a name
    • c() is the combine function that combines elements together
    • install.packages() installs packages
    • library() loads packages you’ve installed so you can use functions and data that are part of that package

Three sources of functions

Three sources of functions:

  • base R
    • <-; mean(x); library("package_name")
  • packages
    • install.packages("packageName)"
    • remotes::intall_github("user/repository")
  • You
    • my_function <- function(x){x^2}

Functions are like recipes

They have:

  • names

  • ingredients (inputs)

  • steps that tell you what to do with the ingredients (statements/code)

  • tasty results from applying those steps to given ingredients (outputs)

(Source)

Can I kick it?

can_x_kick_it <- function(x){
  # Determine if x can kick it
  # If x in A Tribe Called Quest
  if(x %in% c("Q-Tip","Phife Dawg",
              "Ali Shaheed Muhammad", 
              "Jarobi White")){
    return("Yes you can")
  }else{
    return("Before this, did you really know what live was?")
  }

}
can_x_kick_it("Q-Tip")
[1] "Yes you can"
can_x_kick_it("Paul")
[1] "Before this, did you really know what live was?"

Getting setup to work in R

Each time you start a project in R, you will want to:

  • Set your working directory

  • Load (and if needed, install) the R packages you will use

  • Set any “global” options you want

  • Load the data you’ll be using

Set your working directory

Load (and if needed, install) the R packages you will use

  • Install packages once1 with install.packages("package_name")

  • Load packages every session with library("package_name")

Install packages for the lab

Install packages for the lab

Let’s install the tidyverse and COVID19.

  • Create a new code chunk
  • Label it libraries
  • Copy and paste the following into your console
install.packages("tidyverse")
install.packages("COVID19")
  • Once you’ve installed these packages comment out the code
# install.packages("tidyverse")
# install.packages("COVID19")

Keyboard Shortcuts to toggle # comments

macOS: CMD + SHIFT + C

PC: CTRL + SHIFT + C

Loading the tidyverse and COVID19 packages

  • Type the following into your code chunk:
library("tidyverse")
library(COVID19)

Set any “global” options you want

Here are the global options for these slides:1

# Options for these slides
knitr::opts_chunk$set(
  warning = FALSE,       # Don't display warnings
  message = FALSE,       # Don't display messages
  comment = NA,          # No prefix before line of text
  dpi = 300,             # Figure resolution
  fig.align = "center",  # Figure alignment
  out.width = "80%",     # Figure width
  cache = FALSE          # Don't cache code chunks
  )

Load the data you’ll be using

There are three ways to load data.

  • Load a pre-existing dataset
    • data("dataset") will load the dataset named “dataset”
    • data() will list all datasets
  • Load a .Rdata/.rda file using load("dataset.rda")
  • Read data of a different format (.csv, .dta, .spss) into R using specific functions from packages like haven and readr

Working with Data in R

Overview: Working with Data in R

  • Loading data into R

  • Looking at your data

  • Cleaning and transforming your data

Loading data into R

There are three ways to load data.

  • Load a pre-existing dataset
    • data("dataset") will load the dataset named “dataset”
    • data() will list all datasets
    • Useful for tutorials, working through examples/help
  • Load a .Rdata/.rda file using load("dataset.rda")
  • Read data of a different format (.csv, .dta, .spss) into R using specific functions
    • We will use functions from the haven and readr packages to read data from the web and stored locally on your computer

Loading state-level Covid data

Loading state-level Covid data

The code below downloads two years of daily state-level Covid data:

covid <- COVID19::covid19(
  country = "US",
  start = "2020-01-01",
  end = "2022-12-31",
  level = 2,
  verbose = F
    
)

Please run the following1

load(url("https://pols1600.paultesta.org/files/data/covid.rda"))

Loading state-level Covid data

  • country = US tells the function we want data for the US
  • start = "2020-01-01" sets the start date for the data
  • start = "2020-01-01" sets the end date for the data
  • level = 2 tells the function we want state-level data
  • verbose = F tells the function not to print other stuff
covid <- COVID19::covid19(
  country = "US",
  start = "2020-01-01",
  end = "2022-12-31",
  level = 2,
  verbose = F
    
)

Looking at your data

Anytime you load data into R, try some combination of the following to get a high-level overview (HLO) of the data

  • dim(data) gives you the dimensions (# of rows and columns)
  • View(data) opens data in a separate pane
  • print(data); data will display a truncated view of data in your console
  • glimpse(data) will show a transposed (switch columns and rows) version of data with information on variable type
  • head(data) shows you the first 5 rows
  • tail(data) shows you the last 5 rows
  • data$variable extracts variable from data
  • table(data$variable) creates a frequency table
    • Good for categorical data
  • summary(data$variable) summary statistics
    • Good for numeric data

HLOs allow you to

  • Describe the structure of your data:
    • How many observations (rows)
    • How many variables (columns)
  • Describe the unit of analysis
    • What does a row in your data correspond to
  • Identify the class and type of variables (columns)
    • Numeric, character, logical
    • Is there missing data (NAs)
  • Figure out what transformations, cleaning, and recoding you need to do

Data Wrangling in the Tidyverse

The Tidyverse

The tidyverse is an opinionated collection of R packages designed for data science. All packages share an underlying design philosophy, grammar, and data structures.

For more check out R for Data Science

Tidy Data

Tidy data is a standard way of mapping the meaning of a dataset to its structure.

A dataset is messy or tidy depending on how rows, columns and tables are matched up with observations, variables and types. In tidy data:

  • Every column is a variable.

  • Every row is an observation.

  • Every cell is a single value.

More

dplyr functions for data wrangling

Today and this week will begin learning some tools for selecting and transforming data:

  • select() to select columns from a dataframe
  • filter() to select rows from a dataframe when some statement is TRUE
  • mutate() to create new colums
    • case_when() to recode values when some statement is TRUE
  • summarise() to transform many values into one value
  • group_by() to create a grouped table so that other functions are applied separately to each group and then combined

The %>% (“pipe” operator)

The %>% lets us chain functions together so we can read left to right

slice(filter(covid, administrative_area_level_2 == "Rhode Island"), 1)

Becomes

covid %>% 
  filter(administrative_area_level_2 == "Rhode Island") %>% 
  slice(1)

Keyboard Shortcuts for %>%

macOS: CMD + SHIFT + M

PC: CTRL + SHIFT + M

Describing Data in R

Descriptive Statistics

When social scientists talk about descriptive inference, we’re trying to summarize our data and make claims about what’s typical of our data

  • What’s a typical value
    • Measures of central tendency
    • mean, median, mode
  • How do our data vary around typical values
    • Measures of dispersion
    • variance, standard deviation, range, percentile ranges
  • How does variation in one variable relate to variation in another
    • Measures of association
    • covariance, correlation

Using R to Summarize Data

Here are some common ways of summarizing data and how to calculate them with R

Description Usage
sum sum(x)
minimum min(x)
maximum max(x)
range range(x)
mean mean(x)
median median(x)
percentile quantile(x)
variance var(x)
standard deviation sd(x)
rank rank(x)

All of these functions have an argument called na.rm=F. If your data have missing values, you’ll need to set na.rm=F (e.g. mean(x, na.rm=T))

Exploring Covid-19 Data for Lab

Exploring Covid-19 Data for Lab

Let’s spend the rest of class, exploring what seems like a simple question

On average, did states that adopted mask mandates have lower rates of new cases?

Tasks

  1. Get a high level overview of our data

  2. Subset the data to just U.S. States

  3. Recode our data to get a measure of new Covid cases and what face mask policy policy was in place

  4. Summarize the average number of new cases by face mask policy.

1. Get a high level overview of our data

  1. Create a new chunk
  2. Label it #| label:"HLO"
  3. Run the following code
  4. Comment code with #
dim(covid)
head(covid)
glimpse(covid)
table(covid$administrative_area_level_2)
length(unique(covid$administrative_area_level_2))
covid$confirmed[1:10]
covid %>%
  select(administrative_area_level_2, date, confirmed) %>%
  slice(10:20)
summary(covid$confirmed)
table(covid$facial_coverings)

Answer the following

  • How many observations are there (rows)
  • How many variables (columns)
  • What’s the unit of analysis?
    • In words, how would you describe what a row in your data set corresponds to?
  • Are there any missing values for confirmed
  • What range of values can facial_coverings take?1

Subsetting our data to only U.S. States

Goal: Subset our Covid data to include only the 50 states + DC

Steps:

  1. Create a vector of the territories we don’t want

  2. Use the filter() command to “filter” out these territories

1. Create a vector of the territories we don’t want

territories <- c(
  "American Samoa",
  "Guam",
  "Northern Mariana Islands",
  "Puerto Rico",
  "Virgin Islands"
  )

2. Use the filter() command to “filter” out these territories

covid_us <- covid %>%
  filter(!administrative_area_level_2 %in% territories )

dim(covid)
[1] 58809    47
dim(covid_us)
[1] 53678    47

Creating new variables for analysis

  • Goal: We need new variables that describe:

    • the number of new Covid-19 cases on a given date

    • the face mask policy in place

  • Steps:

    • Use mutate(), group_by() and lag() to calculate new_cases from total confirmed cases

    • Use mutate(), case_when() and abs() to turn numeric facial_coverings into categorical factor variable

Calculate new Covid-19 cases

Please run and comment the following code:

dim(covid_us)
[1] 53678    47
covid_us %>%
  mutate(
    state = administrative_area_level_2,
  ) %>%
  dplyr::group_by(state) %>%
  mutate(
    new_cases = confirmed - lag(confirmed),
    new_cases_pc = new_cases/population *100000
    ) -> covid_us
dim(covid_us)
[1] 53678    50

Create Face Mask Policy variable

covid_us %>%
  mutate(
    face_masks = case_when(
      facial_coverings == 0 ~ "No policy",
      abs(facial_coverings) == 1 ~ "Recommended",
      abs(facial_coverings) == 2 ~ "Some requirements",
      abs(facial_coverings) == 3 ~ "Required shared places",
      abs(facial_coverings) == 4 ~ "Required all times",
      
    ) 
  ) -> covid_us

levels(factor(covid_us$face_masks))
[1] "No policy"              "Recommended"            "Required all times"    
[4] "Required shared places" "Some requirements"

Make face_masks a factor to reflect order of policies

levels(factor(covid_us$face_masks))
[1] "No policy"              "Recommended"            "Required all times"    
[4] "Required shared places" "Some requirements"     
covid_us %>%
  mutate(
    face_masks = factor(
      face_masks,
      levels = c(
        "No policy", 
        "Recommended", 
        "Some requirements",
        "Required shared places",
        "Required all times"
        )
    )
  ) -> covid_us

levels(covid_us$face_masks)
[1] "No policy"              "Recommended"            "Some requirements"     
[4] "Required shared places" "Required all times"

Calculate the Average Number of Covid-19 cases by Face Mask Policy

  • Goal: On average, did states that adopted mask mandates have lower rates of new cases?

  • Steps: use filter(), group_by() and summarise() and mean() to calculate the average number of cases for each level of the face_masks policy variable

Face Masks and New Covid-19 Cases (per 100k)

covid_us %>%
  filter(!is.na(face_masks))%>%
  group_by(face_masks)%>%
  summarize(
    `Average No. of New Cases` = round(mean(new_cases_pc, na.rm=T),2)
  )%>%
  rename(
    "Face Mask Policy" = face_masks
  ) -> face_mask_summary

Face Masks and New Covid-19 Cases (per 100k)

  • What should we conclude?

  • What’s wrong with this simple comparison?

  • What’s a better comparison? (Thursday)

Face Mask Policy Average No. of New Cases
No policy 10.26
Recommended 16.61
Some requirements 36.18
Required shared places 29.38
Required all times 32.18

Commented Code

# ---- Libraries ----
## Uncomment to install
# install.packages("tidyverse")
# install.packages("COVID19")

library("tidyverse")
library("COVID19")

# ---- Load data ----
load(url("https://pols1600.paultesta.org/files/data/covid.rda"))

# ---- Subset to US states and DC ----

territories <- c(
  "American Samoa",
  "Guam",
  "Northern Mariana Islands",
  "Puerto Rico",
  "Virgin Islands"
  )


covid_us <- covid %>%
  filter(!administrative_area_level_2 %in% territories )

## Check subsetting
dim(covid)[1] > dim(covid_us)[1]

# ---- Recode covid_us ----

covid_us %>%
  mutate(
    state = administrative_area_level_2,
  ) %>%
  dplyr::group_by(state) %>%
  mutate(
    new_cases = confirmed - lag(confirmed),
    new_cases_pc = new_cases/population *100000
    ) %>%
  mutate(
    face_masks = case_when(
      facial_coverings == 0 ~ "No policy",
      abs(facial_coverings) == 1 ~ "Recommended",
      abs(facial_coverings) == 2 ~ "Some requirements",
      abs(facial_coverings) == 3 ~ "Required shared places",
      abs(facial_coverings) == 4 ~ "Required all times"
      ) 
    ) %>%
  mutate(
    face_masks = factor(
      face_masks,
      levels = c(
        "No policy", 
        "Recommended", 
        "Some requirements",
        "Required shared places",
        "Required all times"
        )
      )
    )-> covid_us


# ---- Calculate new cases per capita by facemask policy
covid_us %>%
  filter(!is.na(face_masks))%>%
  group_by(face_masks)%>%
  summarize(
    `Average No. of New Cases` = round(mean(new_cases_pc, na.rm=T),2)
  )%>%
  rename(
    "Face Mask Policy" = face_masks
  ) -> face_mask_summary

face_mask_summary

Summary

Summary

After today, you should have a better sense of

  • How to write R code using Quarto and R Markdown

  • How to install packages and load libraries

  • Some of different types and shapes of data

  • How to get a high level overview of your data

  • How to transform, recode, and summarise data using dplyr and the tidyverse

  • How describe typical values and variation in data

  • How to explore substantive questions using these these typical values

Congrats!

  • We covered A LOT

  • It’s OK to feel overwhelmed

    • But please don’t suffer in silence

  • Don’t worry if everything didn’t make sense.

    • Eventually it will, but this takes time and practice
    • Testa’s 50-50 rule
    • FAAFO

POLS 1600