1  Introduction to midfield

1.1 which cips have the most time in program?

okay so we’ve decided which cip code set we’re looking at! Great!

Lets see some graphs describing our data within that set. Which cips are the most represented?

Code

most_represented_majors <- term %>% 
  select(mcid, cip6) %>% 
  filter(cip6 %in% cips_of_interest) %>% 
  select(cip6) %>% 
  arrange(cip6) %>% 
  group_by(cip6) %>% 
  summarise(count = n()) %>% 
  arrange(-count) 

mrm_plot <- most_represented_majors %>% 
  ggbarplot(x = "cip6", y = "count") + rotate_x_text(45)
ggexport(mrm_plot, filename = "figures/most_represented_majors.png", verbose = FALSE)
knitr::include_graphics("figures/most_represented_majors.png")

Code

most_represented_major <- most_represented_majors %>% 
  head(1) %>% 
  pull(cip6)

# make a list of students who obtained this degree
students_who_graduated_major <- degree %>% 
  filter(cip6 == most_represented_major) %>% 
  select(mcid) %>% 
  unique() %>% 
  pull(mcid)

# courses taken by students in most represented major
courses_taken <- course %>%
  filter(mcid %in% students_who_graduated_major) %>%
  mutate(an = str_c(abbrev, " ", number)) %>%
  select(an) %>%
  group_by(an) %>% 
  summarise(count = n()) %>%
  arrange(-count)

courses_taken_plot <- courses_taken %>%
  head(20) %>%
  ggbarplot(x = "an", y = "count") + rotate_x_text(45)
ggexport(courses_taken_plot, filename = "figures/courses_taken.png", verbose = FALSE)
knitr::include_graphics("figures/courses_taken.png")

Make a stacked bar plot to break down which relative semester each course was taken

Code

compute_semester_indicies <- function(single_student_record) {
  bar <- single_student_record %>% 
    tibble() %>% 
    mutate(term_course = as_factor(term_course)) %>% 
    mutate(semester_taken = as.integer(term_course)) %>% 
    list()
  return(bar)
}

# this is the continuous interpretation
when_courses_were_taken_discrete <- course %>%
  filter(mcid %in% students_who_graduated_major) %>%
  mutate(course = str_c(abbrev, " ", number), 
         mcid, term_course) %>%
  select(mcid, course, term_course) %>%
  nest_by(mcid) %>% 
  mutate(data = compute_semester_indicies(data)) %>% 
  unnest(cols = c(data)) %>% ungroup(mcid) %>% 
  select(course, semester_taken)

# this summarize converts our interpretation to categorical
when_courses_were_taken <- when_courses_were_taken_discrete %>%
  group_by_all() %>% 
  summarise(COUNT = n()) %>%
  ungroup()

`summarise()` has grouped output by 'course'. You can override using the
`.groups` argument.

Code

total_counts <- when_courses_were_taken %>% 
  group_by(course) %>% 
  summarise(total_count = sum(COUNT)) %>% 
  ungroup()

when_courses_were_taken <- left_join(when_courses_were_taken,
                                       total_counts, by = "course") %>% 
  arrange(-total_count) %>% 
  mutate(semester_taken = as_factor(semester_taken))


when_bio_courses_were_taken <- when_courses_were_taken %>%
  filter(grepl("BIO", course)) 

Code

# most taken courses stacked by what sequence position they were in
when_courses_were_taken_plot <- when_courses_were_taken %>% 
  filter(total_count > 100) %>%
  arrange(-COUNT) %>% 
  head(50) %>% 
  ggbarplot(x = "course", 
            y = "COUNT", 
            fill = "semester_taken") + rotate_x_text(45)
ggexport(when_courses_were_taken_plot,
         filename = "figures/when_courses_were_taken_barplot.png", verbose = FALSE)
knitr::include_graphics("figures/when_courses_were_taken_barplot.png")

Code

# courses with high variation of sequence position
when_courses_were_taken_barplot <- when_courses_were_taken %>%
  select(course) %>%
  group_by(course) %>% 
  summarize(n = n()) %>%
  ungroup() %>% 
  arrange(-n) %>%
  head(20) %>%
  ggbarplot(y = "course", x = "n",
            title = "Number of different sequence positions per course") + rotate_x_text(45)

Storing counts in `nn`, as `n` already present in input
ℹ Use `name = "new_name"` to pick a new name.

Code

ggexport(when_courses_were_taken_barplot, filename = "figures/when_courses_were_taken.png", verbose = FALSE)
knitr::include_graphics("figures/when_courses_were_taken.png")

Code

highly_taken_classes <- total_counts %>% 
  arrange(-total_count) %>% head(10)

when_courses_were_taken_discrete %>% 
  filter(course %in% highly_taken_classes$course & semester_taken < 10) %>%
  ggviolin(x = "course", y = "semester_taken") +
  rotate_x_text(45) + 
  grids() -> p
p <- ggpar(p, yticks.by = 1)
ggexport(p, filename = "figures/when_courses_taken_violin.png", verbose = FALSE)
knitr::include_graphics("figures/when_courses_taken_violin.png")