A description of the fields in the cs1064 WE dataset

cs1064_worked_examples_data-documentation.json

# A list of dictionaries, where each dictionary represents a student and their assignment data
# Each user is uniquely defined by a given name, e.g., "Abomasnow"
# Each assignment's name is the human readable assignment name, e.g., "#43.4) Multiple Returns"
# The "events" key maps to a list of the student's events, given as a list of 4 values.
# The integer timestamp is an epoch time; note that some events trigger at the exact same time.
# Here are some common event/action types:
#   code/set is when a student edits their code
#   editor/run is when a student runs their code
#   editor/reset is when they reset their code, and there's other editor/<Something> actions too
#   feedback/<Something> is the specific kind of feedback that the student received
#   worked_examples/load is when a student loads a worked example
#   worked_examples/position is a batched collection of scroll events (snapshot taken every second if there was a change)
# Just to explain that last event a little more: The format is:
#   "Name of HTML File[[Top of scroll bar, Height of document, Height of window, Whether the page has focus], ...]"
# So this:
#   "pet_count.html[[193,2375,806,true],[195,2375,806,true]]"
# Means the "pet_count.html" file, with two scroll events, 
#    and the user scrolled 193/2375 of the way down the page followed by 195/2375 a second later.

[
  {
    "user": {
      "section": "One of 'Fall 2017', 'Bart - MWF', or 'Panagiotis - TR', indicating last fall or one of the two spring sections.",
      "user_id": "A unique user ID that identifies the student,",
      "name": "A human-friendly, coded name for the student (unique).",
    },
    "assignment": {
      "<Assignment Name>": {
        "correct": "Either true or false, depending on if they finished the problem.",
        "attempted": "Either true or false, depending on if they even tried the problem.",
        "runs": "The number of times they ran their code, regardless of whether it worked.",
        "time_on_task": "The number of seconds they spent working on the problem. Gaps of 45+ seconds were removed when calculating this.",
        "version": "The number of times they edited their code.",
        "events": [
          ["<Integer Timestamp>", "<Event Type>", "<Event Action>", "<Event Body>"],
          ...
          ],
      }
  }, ...
]