You can now download the data set we have prepared for you and put it into your project folder:
The data set is from the package palmerpenguins (v0.1.1) and contains the recorded bill lengths and sex of penguins living on three islands in the Palmer Archipelago, Antarctica. It was made available by Allison Horst, Alison Hill, and Kristen Gorman under the license CC0 1.0.
Everything you put into the project folder will be shared publicly. For reasons of reproducibility, this should include the data you analyze. Of course, you should only share them to the extent you are allowed to. Besides copyright and similar rights, you need to take into account applicable privacy laws (e.g., the GDPR for European citizens) and contractual obligations (e.g., with your data provider).
Privacy laws and contractual obligations may require you to create an anonymized or synthetic data set1, in which case you should provide a reference to a repository where the originally measured data can be obtained from. For further information, you can watch the talk “Data anonymity” by Felix Schönbrodt recorded during the LMU Open Science Center Summer School 2023 and have a look at the accompanying slides. Another resource to look at is the presentation “Creating reproducible packages when data are confidential” by Vilhuber (2024).
Whether or not distributing the data set, it is important to document the meaning (e.g., units) and values of its variables. This is typically done with a data dictionary (also called a codebook).
The recommendation for data dictionaries vary between fields - both in terms of the recommended content (i.e., what exactly should be documented) and the technical implementation (i.e., which file formats should be used).
For the purpose of this exercise, we keep it easy and propose that you manually create a file with the data dictionary (e.g., as a table in .xlsx, .docx, .ods, or as a Markdown table), documenting only the bare minimum.
Our in-depth supplementary material “Automatic Generation of Data Dictionaries” explains how you can automatically create a data dictionary with an R package. The package reads the data set and extracts relevant information from it. This approach is in particular useful if you need to document data sets with many variables.
This advanced chapter also contains a section on how to create machine-readable data dictionaries.
Most standards for data dictionaries require at least this information for each variable in your data set:
name: The (machine‑readable) name of the variablelabel: A short, human‑readable title or label for the variabletype: The data type of the variable (e.g., integer, float, string, date)description: A brief description of what the variable measures or representsvalues (for categorical variables): A mapping of codes to their meanings, for example: 1 = Male, 2 = Female, 9 = Missing)units (for numeric measures): The units of measurement (e.g., kg, USD, years, or the scale of a survey response),missing_codes: Any special codes used to denote missing or non‑applicable values (e.g., -99 = Not answered)Here’s an example from a different data set, with variables in rows, and the dictionary in columns:
| name | label | type | description | values | units | missing_codes |
|---|---|---|---|---|---|---|
| gender | Gender | integer | self‑identified gender | 1 = Male; 2 = Female; 3 = Other; 9 = Missing | 9 = Missing | |
| age | Age | integer | age in years | years | -99 = Not answered | |
| blood_pressure | Blood Pressure (systolic) | integer | systolic blood pressure | mmHg | -99 = Not measured | |
| life_satisfaction | Life Satisfaction | integer | “How satisfied are you with your life?” | 1 = Very dissatisfied; 2 = Dissatisfied; 3 = Neutral; 4 = Satisfied; 5 = Very satisfied | scale (1–5) | -99 = Not answered |
Now go ahead and create a data dictionary for the penguins data set, in a software (text or spreadsheet) of your choice.
Save the data dictionary in the same folder as the actual data set file.
| name | label | type | description | values | units | missing_codes |
|---|---|---|---|---|---|---|
| species | Species | string | Penguin species | Adelie; Gentoo; Chinstrap | NA = Missing | |
| island | Island | string | Island where individual was observed | Torgersen; Biscoe; Dream | NA = Missing | |
| bill_length_mm | Bill length | float | Length of the bill (beak) | mm | NA = Missing | |
| bill_depth_mm | Bill depth | float | Depth of the bill (beak) | mm | NA = Missing | |
| flipper_length_mm | Flipper length | integer | Length of the flipper | mm | NA = Missing | |
| body_mass_g | Body mass | integer | Body mass | g | NA = Missing | |
| sex | Sex | string | Sex of the penguin | male; female | NA = Missing | |
| year | Year | integer | Year of data collection | 2007; 2008; 2009 |
All data relied upon should be cited in the manuscript to allow for precise identification and access. Now, it’s your turn to add an appropriate citation for the data set to the manuscript.
Hints:
citation()2.Show the correct reference of the data set:
citation("palmerpenguins", auto = TRUE) |>
transform(key = "horst2020") |>
toBibtex()Bibliography.bib
@Manual{horst2020,
title = {palmerpenguins: Palmer Archipelago (Antarctica) Penguin Data},
author = {Allison Horst and Alison Hill and Kristen Gorman},
year = {2022},
note = {R package version 0.1.1},
url = {https://CRAN.R-project.org/package=palmerpenguins},
doi = {10.32614/CRAN.package.palmerpenguins},
}Copy the BibTeX entry to the file Bibliography.bib. Then, find the line in the manuscript that says “cite data here” and replace it with a sentence such as the following:
Manuscript.qmd
The analyzed data are by @horst2020.Render the document to check that the citation is displayed properly.
Terminal
quarto render Manuscript.qmdCongrats! You documented your data set and cited it correctly.
To finalize this step, you can go through the commit routine:
Terminal
git status
git add .
git commit -m "Add data"