Datahub/DSMLP Assignment Grading Tools


Overview: Grading Assignments


Formgrader/nbgrader: we continue to support for assignment creation, distribution, and auto-grading.

Otter-Grader is a new auto-grader available to instructors/TAs that enables grades to be uploaded to your course assignments in Canvas, via Gradescope. (Otter-Grader Gradescope documentation)  

Please see the grading tool comparison table for more information. IT Service staff are available for consultation on grading tool selection/help, as well as putting you in contact with other instructors about which grading tools they use.

Distinction \ Product

Gradescope Autograder (vanilla)

Otter-grader using Gradescope Autograder 

nbgrader

Service Provider

Gradescope

Gradescope

DataHub

Support Provider

Gradescope

Gradescope / community

ETS

Documentation

https://gradescope-autograders.readthedocs.io/en/latest/

https://otter-grader.readthedocs.io/en/latest/

https://nbgrader.readthedocs.io/en/latest/

Assignment and grading user interface in datahub

No

Yes

Assignment Creation

Gradescope

DataHub

Custom Grading environment

Docker image built by Gradescope with instructor customization

Docker image built by Gradescope with instructor customization

Docker image built by ETS

Python unit-test

Yes

Yes, can be done within cells

Jupyter Notebook unit-test

No

Yes (generated by otter-grader)

Yes

Other programming languages unit-tests (C/C#, Java, MySQL, etc)

Native support

No

No

Instructor unit-test development

Local environment through code

Local environment through code (okpy templates)

Native support (web interface / Jupyter)

Unit-tests run on

Gradescope’s infrastructure on AWS

DataHub cluster

Manual grading

(e.g. short-response)

No (but could run as a separate assignment)

Yes (PDF generation)

Yes (web/feedback interface)

Manual grading happens on

Gradescope (separate non-autograder assignment, manual)

Gradescope (separate non-autograder assignment, automatic)

DataHub

Instructor Feedback

Only for (non-autograder assignment)

Yes (use manual grading)

Student Feedback before submission

Test-run with public tests

Validate (public/visible assert cells)

Student Feedback after grades release

Score breakdown with visibility customization; Email

Score breakdown in notebook for assert cells

Location where students complete assignments  

Anywhere (usually a local setup)

Anywhere (usually a local setup or DataHub)

DataHub

What do students submit

Code artifact through Gradescope

Jupyter notebook through Gradescope

Jupyter notebook through DataHub

Canvas Course Integration

Yes (by Gradescope)

Yes (by ETS)

Grades Import to Canvas

Yes, via Gradescope

Yes, manually and automatically (see Nb2Canvas)



If you still have questions or need additional assistance, please email datahub@ucsd.edu or visit support.ucsd.edu.