With the goal of learning my students’ names faster, I wrote an R script that will convert a roster (“People” in a Canvas course) to a gt table for later offline use.

knitr::opts_chunk$set(eval = FALSE)

Setup

In your Canvas course, go to People, right-click on the roster, and click “Save Page As”

be sure to save the “complete” web page (this creates a folder with all of the students’ pictures)
place the also downloaded HTML file into the folder
place this code script into the folder

# change this file name as needed
HTML_file_name <- "Course Roster S23-BIO 018 01.htm"

library("gt")
library("gtExtras")
library("rvest")
library("tidyverse")

Roster table

https://www.r-bloggers.com/2020/04/scrape-html-table-using-rvest/

content <- read_html(HTML_file_name)
tables <- content |> html_table(fill = TRUE)
roster_table <- tables[[1]]

roster_df <- roster_table |>
  select(Name, Section) |>
  mutate(discussion_tag = str_extract(Section, "Discussion-\\d\\dD")) |>
  separate(discussion_tag, sep = "-", into = c("disregard", "discussion_section")) |>
  select(Name, discussion_section) |>
  
  # remove instructor of record
  filter(!is.na(discussion_section))

Images

https://community.rstudio.com/t/scraping-images-from-the-web/133239

image_urls <- content |> html_elements("img")
num_images <- length(image_urls)

image_df_raw <- data.frame(html_raw = rep(NA, num_images))
for(img in 1:num_images){
  image_df_raw$html_raw[img] <- as.character(image_urls[[img]])
}

image_df <- image_df_raw |>
  
  # skip first two images?
  slice(3:n()) |>
  
  # roundabout because some people's names have more than 2 words
  separate(col = "html_raw", sep = "alt=", 
           into = c("part1", "part2")) |>
  
  separate(col = "part1", sep = " ", into = c("tag1", "src")) |>
  separate(col = "src", sep = "/", into = c("stem", "profile_picture")) |>
  separate(col = "part2", sep = " aria", into = c("student_name", "tag2")) |>
  
  select(profile_picture, student_name)

# remove quotation marks
image_df$profile_picture <- str_replace_all(image_df$profile_picture, "\"", "")
image_df$student_name <- str_replace_all(image_df$student_name, "\"", "")

Merge

roster_df <- roster_df |>
  left_join(image_df, by = c("Name" = "student_name")) |>
  select(profile_picture, Name, discussion_section)

gt

https://jthomasmock.github.io/gtExtras/reference/gt_img_rows.html

roster_gt <- roster_df |>
  gt() |>
  gt_img_rows(columns = profile_picture, img_source = "local")

gtsave(roster_gt, "roster.png")