Randomly assign students into seminar reading groups of equal sizes, different groups each week

random-group-assignment.R

# get canvas gradebook as CSV...

library(tidyverse)
students <-
  readr::read_csv("2023-10-01T2053_Grades-ARCHY_482_A.csv") %>%
  # exclude test student
  filter(!is.na(`SIS User ID`)) %>%
  select(Student) %>%
  as.data.frame()

# add one column per seminar to hold the group assignments
seminar_names <- str_glue('seminar_{1:10}_reading_group')
students[, seminar_names] <- NA

# randomly allocate students into n groups
x <- 1:nrow(students) # how many students do we have?
# we have this many students
length(x)

# if we want 4 students per group, how many groups? 
students_per_group <- 4
n <- length(x) / students_per_group
group_numbers <- cut_number(x, n, labels = FALSE)

# make a dataframe of group assignments for each seminar
students_seminar_groupings <-
  students %>%
  mutate(across(starts_with('seminar'),
                ~sample(group_numbers)))

# check that each group has about the same number of people each week
checks <-
  map_dfc(seminar_names,
          ~students_seminar_groupings %>%
            group_by(!!rlang::sym(.x)) %>%
            tally()
  )

# write to CSV
write_csv(students_seminar_groupings,
          "students_seminar_groupings.csv")

# for week by week, two groups per seminar: 
# 1. sort by reading group number (Data -> sort by range)
# 2. create a column "seminar_1_mixed_group"
# 3. in the 'seminar_1_mixed_group' column, enter a repeating sequence of 1 -> n, where n is the number of groups (in this case, 4)
# 4. sort by student name so they can easily find their numbers by browsing
# 5. apply formatting to get alternating row colours (Format -> alternating colours)

# how to randomly assign discussion group leaders for each group?

seminar_reading_assignments <- 
  vector("list", length = length(seminar_names))
names(seminar_reading_assignments) <- seminar_names

for(i in seminar_names){
  seminar_reading_assignments[[i]] <- 
  students_seminar_groupings %>% 
    select(Student, !!rlang::sym(i)) %>% 
    arrange(!!rlang::sym(i)) %>% 
    group_by(!!rlang::sym(i)) %>%       
    mutate(!!paste0(i, "_leader") := replace(!!rlang::sym(i), 
                                             sample(1:students_per_group, 1),  
                                     paste0(!!rlang::sym(i), 
                                            "L"))) %>% 
    mutate(seminar_mixed_group = 1:n()) %>% 
    group_by(seminar_mixed_group) %>%  
    arrange(seminar_mixed_group) %>% 
    mutate(seminar_mixed_group_leader = replace(seminar_mixed_group, 
                                                sample(1:students_per_group, 1), 
                                              paste0(seminar_mixed_group, "L"))) %>% 
    arrange(Student) %>% 
    ungroup() %>% 
    dplyr::select(-c(seminar_mixed_group,
           !!rlang::sym(i))) 
}

map2(seminar_reading_assignments,
    seminar_names,
    ~write_csv(.x,
               file = paste0(.y, ".csv")))

# check to see if the distribution is roughly fair

map2(seminar_reading_assignments, 
     paste0(seminar_names, "_leader"), 
    ~ rename(all_of(.), "seminar_reading_group_leader" = .y)) %>% 
  bind_rows(., .id = "seminar")  %>% 
  mutate(leader = case_when(
    str_detect(seminar_reading_group_leader, "L") ~ TRUE, 
    str_detect(seminar_mixed_group_leader, "L") ~ TRUE
   ))  %>% 
  group_by(Student) %>% 
  tally(leader) %>% 
  ggplot() +
  aes(Student, n) +
  geom_col() +
  coord_flip()