R for Everything: Getting, Cleaning, Visualizing, and Analyzing Data

Philosophy

1. Data should be in a consistent form (data frame)
2. Each variable is a column
3. Sequence clear, logical steps

Installation

install.packages("tidyverse")
library(tidyverse) #loads readr, dplyr, tidyr, purrr, ggplot2, tibble 

Purpose

Fast and friendly way to read rectangular data.

Usage

df = data_frame(x = 1:3, y = x^2, z = sample(letters, 3, replace = TRUE))
write_csv(df, "my_csv.csv")
read_csv("my_csv.csv")

# A tibble: 3 × 3
      x     y     z
  <int> <dbl> <chr>
1     1     1     v
2     2     4     z
3     3     9     i

select(iris, Sepal.Length, Petal.Length, Species)

filter(iris, Sepal.Length > 7)

mutate(iris, sepal = Sepal.Length + Sepal.Width)

summarize(iris, avg = mean(Sepal.Length))

The pipe (%>%) chains tidyverse functions together:

iris %>% 
  filter(Sepal.Length > 4) %>% 
  group_by(Species) %>% 
  summarize(avg = mean(Sepal.Width)) %>% 
  arrange(desc(avg))

# A tibble: 3 × 2
     Species   avg
      <fctr> <dbl>
1     setosa 3.428
2  virginica 2.974
3 versicolor 2.770

Purpose

Package to tidy and reshape data.

Usage

table4 %>% gather(year, cases, -country)

Purpose

Work with lists and facilitate iteration.

Usage

map_dbl(1:3, log, base = 2)

[1] 0.000000 1.000000 1.584963

Using purrr to calculate Ed-Data's Ethnic Diversity Index (EDI)

edi <- function(df) {
  if (!is.data.frame(df)) stop("student_df must be a data frame")
  if (!"ethnicity" %in% names(df)) stop("ethnicity must be a column")
  ur <- c("Decline/Don't Know", "Other", "")
  ur_fraction <- sum(df$ethnicity %in% ur) /
                 sum(!df$ethnicity %in% ur)
  diversity_rating <- df %>% 
    filter(!ethnicity %in% ur) %>% 
    split(.$ethnicity) %>% 
    map(~ nrow(.)/nrow(df)/(1 - ur_fraction)) %>% 
    map_dbl(~ (. - (1/13))^2) %>%  #There are thirteen reported ethnicities
    sum(.) %>% 
    sqrt(.)
  c2 <- -100 * sqrt(13*(13-1))/(13-1)
  100 + (c2 * diversity_rating)
}

Purpose

Data frames with nicer behavior around printing and subsetting

Usage

df1 <- tibble(x1 = 1:3, y1 = 1, z1 = x1 ^ 2 + y1)
df1

# A tibble: 3 × 3
     x1    y1    z1
  <int> <dbl> <dbl>
1     1     1     2
2     2     1     5
3     3     1    10

df1$x

Warning: Unknown column 'x'

NULL

library(rcanvas)

premium_courses <- get_course_list() %>% 
  filter(grepl("Premium", name))

get_emails_and_grades <- function(id) {
  emails <- get_course_items(id, "users", include = "email") %>% 
    select(name, sis_user_id, sis_login_id, email)
  grades <- get_course_items(id, "enrollments") %>% 
    filter(enrollment_state == "active") %>% 
    select(id, user.name, user.sis_user_id, user.sis_login_id, grades.current_score) %>% 
  left_join(emails, grades, by = c("sis_user_id" = "user.sis_user_id"))
}

safe_function <- safely(get_emails_and_grades)

student_data <- premium_courses$id %>% 
  map(safe_function) %>% 
  bind_rows() %>% 
  left_join(premium_courses, by = c("course_id.x" = "id")) %>%
  select(name, sis_user_id, email, course_id.x, grades.current_score, sis_course_id)

student_contact <- read_csv("student_contact.csv") %>% 
  select(sis_user_id = `Student ID`, Student, Question, Answer, email = Email) %>%
  spread(Question, Answer)

# A tibble: 3 × 5
  sis_user_id        Student                  Email   `Parent Email`
        <chr>          <chr>                  <chr>            <chr>
1    A0004325 David Ranzolin  dranzolin@ucscout.org info@ucscout.org
2    A0004375   Sajira Awang     sawang@ucscout.org info@ucscout.org
3    A0004925 Lisa Dominguez ldominguez@ucscout.org info@ucscout.org
# ... with 1 more variables: `Counselor Email` <chr>

email_df <- student_contact %>% 
  left_join(student_data, by = "sis_user_id") %>% 
  filter(grades.current_score < 80) %>% 
  select(-Sections, -course_id.x) %>% 
  rename(counselor_email = `Counselor Email`,
         parent_email = `Parent Email`)

subject <- "UC Scout Weekly Grade Update"
email_sender <- 'UC Scout <info@ucscout.org>' 
body <- "Dear %s,

We're writing to inform you that your current grade in %s is %s. You can view your course progress in your Online Classroom (classroom.ucscout.org). 
Please let us know if you need any assistance. You can also contact your teacher with any further questions or concerns.

Best wishes,
The Scout Team"

email_df2 <- email_df %>%
  mutate(
    To = sprintf('%s <%s>, <%s>, <%s>', name, email, counselor_email, parent_email),
    From = email_sender,
    Subject = subject,
    Body = sprintf(body, name, sis_course_id, grades.current_score)) %>%
  select(To, From, Subject, Body)

library(gmailr)
emails <- email_df2 %>% 
  pmap(mime)

use_secret_file("client_secret_PROJ-NAME.json")

safe_send_message <- safely(send_message)
sent_mail <- emails %>% 
  map(safe_send_message)