roarutility

The goal of roarutility is to process ROAR data. It provides convenience functions to make some common cleaning and processing tasks with ROAR data a bit easier.

Installation

Install the package, once you have access to the roarutility repository, by running the following code.

# install.packages("pak")
pak::pak("yeatmanlab/roarutility")
library(roarutility)

Usage

roar.read.csv

This is a basic example which shows you how to read in ROAR data and remove opt-outs in one line of code.

library(roarutility)
new_data <- roar.read.csv("all_runs.csv", 
              "~/Documents",
              "google.drink.link")

Notice how the output dataframe has removed all possible opt-outs from the most up-to-date opt-out CSV.

clean_strings

This is a basic example which shows you how clean_strings() takes in data and outputs data that has removed extra characters from assigning organization variables and converted empty strings to NA values.

library(roaryutility)
test_df <- data.frame(
  assigning_schools = c("[irNgj3c]", "irNgj3c"),
  age = c("", "6.7")
)

clean_df <- clean_strings(test_df) 
clean_df 
#   assigning_schools  age
# 1           irNgj3c <NA>
# 2           irNgj3c  6.7

Notice how the output dataframe has removed the “[]” characters from the assigning_schools variable and has converted the empty string value in age to an NA value.

remove_empty_cols

This is a basic example which shows you how remove_empty_cols() takes in a dataframe with a column with all NA values and outputs data that has removed the columns with all NA values.

library(roarutility)
test_df <- data.frame(
  firstname = c("Jane", "John", NA, "Kelly"),
  lastname = c("Doe", NA, NA, "Smith"),
  middlename = c(NA, NA, NA, NA)
)

clean_df <- remove_empty_cols(test_df) 
names(clean_df) 
# [1] "firstname" "lastname" 

Notice how the output dataframe has removed the column “middlename” because it consisted of all NA values.

remove_duplicates

This is a basic example which shows you how remove_duplicates() removes all identical rows across every column.

library(roarutility) 
test_df <- data.frame(
  assessment_pid = c("123", "456", NA, "789", "123"),
  roarScore = c(45, 32, 34, 10, 45)
)

clean_df <- remove_duplicates(test_df)
clean_df$assessment_pid
#[1] "123" "456" NA    "789"

Notice how the last row test_df[5,] was removed because it had identical values across both columns (assessment_pid and roarScore) as the first row test_df[1,].

remove_accounts

Removes all indicated accounts from the dataframe. Reseearchers can read in data and indicate which or all of the following types of accounts they would like to remove from the dataframe. The function defaults to removing test, demo, pilot, and QA accounts and defaults to not removing NA assessment_pid. The function runs through the organization IDs (i.e., assigning_districts, etc.). It also uses string detection to determine if there are any “test”, “pilot”, “qa”, or “demo” strings within the assessment_pid column. Finally, it runs through to determine if there are test or demo using the variables is_test_data and is_demo_data (if these accounts were chosen to be removed). If selected, the function will also remove assessment_pid = NA.

estimate_grade

This is a basic example which shows you how estimate_grade() uses age_months_at_run to estimate grade for students this school year and for students who were missing grade values.

library(roarutility)
test_df <- data.frame(
  age_months = c(75, 83, 99, 200),
  user_grade = c(NA, "2", "2", "10"),
  time_started = c("2025-03-12", "2024-02-17", "2026-01-09", "2025-04-19")
)

clean_df <- test_df %>% mutate(user_grade = case_when(
  time_started < as.Date("2024-07-31") ~ map_chr(age_months, estimate_grade),
  TRUE ~ user_grade))
clean_df
#   age_months user_grade time_started
# 1         75       <NA>   2025-03-12
# 2         83          1   2024-02-17
# 3         99          2   2026-01-09
# 4        200         10   2025-04-19

clean_df <- clean_df %>% mutate(user_grade = case_when(
  is.na(user_grade) ~ map_chr(age_months, estimate_grade),
  TRUE ~ user_grade))
#   age_months user_grade time_started
# 1         75          1   2025-03-12
# 2         83          1   2024-02-17
# 3         99          2   2026-01-09
# 4        200         10   2025-04-19

Notice in the first example of estimate_grade(), we use it to adjust grades that were logged prior to the 24-25 school year. The only grade that changes is row 2 and the function accurately recategorizes the grade from 2nd to 1st grade based on the age in the months at the time of the run. In the second example, we use the function to fill in missing grades using age in months. The only grade that changes is in row 1 and function accurately fills in the missing value for grade with “1” for 1st grade.

standardize_grade

This is a basic example which shows you how standardize_grade() uses a grade variable and dataframe to create uniform values in grade.

library(roarutility)
test_df <- data.frame(user.grade = c("2", "1", "01", "2nd", "k",
                                     "Kindergarten", "1", "09"))
clean_df <- standardize_grade(test_df, "user.grade")
clean_df$user.grade
# [1] "2"            "1"            "1"           
# [4] "2"            "Kindergarten" "Kindergarten"
# [7] "1"            "9"  

Notice how the grades went from nonuniform values “2”, “2nd”, “01” to more uniform values which can help researchers with filtering, faceting, and overall data organizations.

filter_assessments

This is a basic example which shows you how filter_assessments maintains the assessments that have completed runs in the first example and completed, best, and reliable runs in the second example.

library(roarutility)
test_df <- data.frame(
  task_id = c("roam-alpaca", "swr", "sre", "letter", "sre-es", "swr-es"),
  completed = c("true", "true", "false", "true", "false", "false"),
  best_run = c(NA, "true", "false", "true", NA, NA),
  reliable = c(NA, "true", "false", "true", NA, NA)
)

clean_df <- filter_assessments(test_df) 
clean_df
#       task_id completed best_run reliable
# 1 roam-alpaca      true     <NA>     <NA>
# 2         swr      true     true     true
# 3      letter      true     true     true

clean_df <- filter_assessments(test_df, completed=TRUE, best_run=TRUE, reliable=TRUE)
clean_df
#   task_id completed best_run reliable
# 1     swr      true     true     true
# 2  letter      true     true     true

Notice how in the first example, filter_assessments() only keeps the assessments where completed==“true”, but did not consider the values for best_run and reliable. In the second example, we indicate that we also want to consider best_run and reliable variables as well as completed. As you can see, the function only keeps those which have all “true” values.