Semester project

Instructions for graduate students

The goal for this project is for you to implement a complete open–source reproduction of an important research paper from your dissertation field. The general steps to completing this project are as follows:

  1. Identify a published paper whose key results you would like to reproduce, and whose reproduction seems feasible.1. You may also choose a paper that has not yet been published, provided you have access to the raw data, understand the authors’ intentions, and a sense of their (preliminary) results. Please consult with Gaurav first if you wish to pursue this path.

  2. Secure the dataset generated in the original study, either directly from the paper or by contacting the authors.

  3. Reproduce, to the best of your abilities, the analyses and figures presented in the original publication. Your work on this reproduction analysis should be tracked with version control, and the final report should be in the form of a qmd document.

  4. Write a report that describes (i) the motivation, methods, and results of the original publication; (ii) the steps you took to reproduce these results, including any problems that got in the way of your reproduction; (iii) what you learned from the reproduction exercise.

  5. Make a presentation, using Quarto, to share your findings with your classmates.

Instructions for undergraduate students

Undergraduate students will choose an openly available dataset from the lterdatasampler project. Your task is to create two non-trivial visualizations2 As above, undergraduate students will write a report that describes (i) the motivation and methods that went into collecting the original dataset; (ii) the steps you took to generate your figures, including any problems that got in the way; (iii) what you learned from the exercise.

Grading

Your final grade will be determined as follows:

Graduate students

  • Proposal presentation (~10 minutes each; Nov. 3rd and 5th): 25 points
  • Final project presentation (15 minutes each with 5 for questions; Week 15 (Dec 1/3)): 25 points
  • Final writeup (Dec 8): 30 points
  • Participation and engagement: 20 points

Undergraduate students

  • Proposal writeup (~10 minutes each; Nov. 3rd and 5th): 15 points
  • Final writeup (Dec 10): 25 points
  • Participation and engagement: 10 points

Further details for each will be provided as the semester progresses.

Footnotes

  1. We will discuss characteristics of “good” paper to reproduce in class. Generally, this means that the data are available (or made available by the authors), but the code is not easily available.↩︎

  2. Here, I take “non-trivial” to mean something more than just plotting data from two existing columns against one another.↩︎