Assignment 1 feedback#

See also the grading rubric.

Follow instructions#

I didn’t penalise reports for being overly long or short. I did give lower scores for code not submitted through the github repository, or for not having the expected selection of sections.

Code#

Code ranged from long notebooks with minimal structure to code being highly functionalised and routines separated into *.py files. Comments on codes ranged from almost nothing, to some comments on individual lines (suitable for reading ones own code), to more detailed comments and markdown text sections between code blocks.

For assignment 2:

  • At a minimum, for those of you working in python notebooks, use markdown text blocks between code cells with some explanation of the following sections. Include some comment to explain each function.

  • We did not teach about commenting and documenting your code in class, so the bar is low. Please see external resources, example here.

  • Try to put multiply-used code into functions for reuse. If you haven’t used functions before, for assignment 2, please try to write at least one function. See examples here (external).

Report#

Reports were all generally good, though some suffered a bit from being a blend between scientific findings and code documentation. Also for this reason, for assignment 2 please submit your report separately as a *.pdf and not as your *.ipynb.

Note that your report can and should be brief (concise), but should give a good overview of the work you carried out. It should not normally refer heavily to the individual coding elements (i.e., does not need to refer to functions used or variables loaded) but should instead focus on the content / what you found from the data. Several reports were a blend between scientific findings and code documentation.

Please do proofread your report. Spelling mistakes detract from the quality of the work.

Content#

A comment on what set apart the very good/exceptional from the solid standard:

  • Intro/background: Included an explanation on why this is worth doing? (I.e., not just because you have to for the class, but by doing it, what are you hoping to find out about the data?)

  • Methods: Justify choices and with enough information for me to see that the reasoning was clear to the author. This included choices about binning and about reference levels.

  • Results: The best examples used their key 1–3 figures to demonstrate a quantitative result, rather than only plotting the data (i.e., a derived quantity, not just time series). The choice of derived quantity is important–what calculation will demonstrate your result most clearly? Plotting the data well is a solid and necessary start; quantifying what you want the reader to see is necessary to get the highest marks. Also, explaining in written form what the main results figures show is important; don’t simply let the figures speak for themselves.

  • Conclusions: Draw conclusions. Don’t leave the reader to draw their own. Be bold and make a claim. (As long as it was backed up by evidence in the results section that you can point at.)