Gradebooks

The main object in gradelib is the Gradebook. It stores all of the data about a course’s grades, and provides basic methods for manipulating that data.

You normally don’t need to create a Gradebook from scratch; instead, you will typically import grades from Gradescope or Canvas using one of the function described in the next section. But it can be useful to know a little about how the Gradebook is structured, in order to have a sense of how gradelib works, and in case you want to implement a grading strategy that isn’t in this package.

You can think of a Gradebook as being comprised of four “core” pieces of information:

• Gradebook.points_earned is a pandas DataFrame containing the points earned by each student on each assignment.

• Gradebook.points_possible is a pandas Series containing the number of points possible on each assignment.

• Gradebook.lateness is a pandas DataFrame measuring how late each student’s submission was for each assignment.

• Gradebook.dropped is a pandas DataFrame indicating whether a student’s submission was dropped for a given assignment.

These four attributes contain all of the information about a student’s performance on each assignment. Together with a fifth attribute, Gradebook.grading_groups, which determines how the assignments should be grouped together for grading purposes, these are sufficient to compute a student’s overall grade in the course.

Beyond the “core” data attributes above, a Gradebook also contains many “derived” attributes, which are computed from the core data. For example, Gradebook.score is a pandas DataFrame containing the score earned by each student on each assignment, which is computed by dividing Gradebook.points_earned by Gradebook.points_possible. The derived attributes are all computed dynamically from the core data.

In gradelib, the grading process amounts to starting with a “raw” gradebook and making a sequence of modifications to the core attributes that reflect the course’s grading policies. For example, you might drop the lowest homework score and penalize late assignments; this is done by modifying the Gradebook.dropped and Gradebook.points_earned attributes, respectively. gradelib provides a number of functions for making these modifications without having to worry about the details of how the core data are structured, but you can also modify the core data directly if you want to implement a grading policy that isn’t supported.