The homework is evaluated concerning these perspectives:
- [P1]: Detailed description of your work (way of thinking, explaining results, etc.)
- [P2]: Quality in terms of readability and efficiency of code (including code comments)
- [P3]: Correctness of the results
Specifically, you will be assigned a score between 1 and 5 for each aspect.
The TAs will evaluate all the homework.
- For each homework, each student will review the assignment of 2 random groups (peer review)
- Therefore, each group will get around 6-8 reviews
The two pieces homework you should review will be communicated through email. Once you have completed your review, you will submit it through a form (submit each repository review in a different form) where you will be asked to tell us the group score for each section and explain your choices.
Since how you do the Peer evaluation will be part of your final grade, here are some examples of good reviews:
Positive
[P1]: Comments are really clear and still manage to be short.
[P2]: Code is really well written: concise and efficient (see the matching names function as an example; really smart and short)
[P3]: Results obtained are clear and well illustrated with very communicative and effective plots.
Negative
[P1]: The textual description for the first two tasks is minimal, but it properly describes the problem. I would hand in a 4.5 for those, but there are only a few comments for the third task. There's commented code left in the modules and even a TODO left in there. One of the members likely didn't finish his task properly, but it is the team's job to at least proofread the final version.
[P2]: For the code quality, utilization of libraries is done throughout, and the code is concise in what it does. One annoyance is that head() or tail() wasn't used, and the outputs are often longer than they could be to show the result format.
[P3]: For the results, first task, there's a clear error with the number of new cases in Liberia for the last period... passing from ~2 new patients a month to ~2k should raise red flags. At least explain why they kept it, especially in a data-wrangling task. In task two, they didn't re-index, split the scientific classification string, or even clean it. It is clear that they went with the solution with the minimal possible effort. In the third task, it is hard to know exactly what was done in each case because of a lack of comments, but the result seems good overall.
Examples of bad reviews:
Good job
Something is missing
The overall review of the homework should (if possible) contain at least two positive and two negative aspects of the assignment you are evaluating.