An Open Dataset Example You Can Turn Into a Public Table
Use the Palmer Penguins CSV to practice documenting units, preserving missing values, checking row counts, and publishing a useful public dataset page.
Use the Palmer Penguins dataset to practice turning an open CSV into a documented public table. The input is a small file of penguin measurements; the result should let readers filter by species, compare measurements, and understand what missing values mean without opening a spreadsheet application.
The simplified dataset has 344 rows and eight columns, covering three penguin species on three islands in Antarctica’s Palmer Archipelago. Each row represents a penguin in the study. Its measurements come from study years 2007–2009, not current monitoring. The package’s dataset reference defines the fields and identifies the original data records.
What makes this an open dataset?
An open dataset is more than a file available to download. Under the Open Definition, anyone must be able to freely access, use, modify, and share it for any purpose, subject at most to conditions preserving provenance and openness.
The Palmer Penguins project states that its data are available under CC0 and credits Dr. Kristen Gorman and Palmer Station Antarctica LTER as the data collectors and providers. CC0 permits copying, modifying, and distributing a work without asking permission, including for commercial purposes.
Separately, the project asks people interested in publishing the data to contact Dr. Gorman about analysis and collaboration on final products. Keep that data-use guidance with your source notes.
A public web page makes a dataset easier to explore. It does not, by itself, establish permission to reuse the underlying data.
Download the CSV and inspect its structure
Save the project’s simplified penguins.csv file. You do not need R to work with this CSV.
These are its first four records, reproduced from the source—not invented sample measurements:
species,island,bill_length_mm,bill_depth_mm,flipper_length_mm,body_mass_g,sex,year
Adelie,Torgersen,39.1,18.7,181,3750,male,2007
Adelie,Torgersen,39.5,17.4,186,3800,female,2007
Adelie,Torgersen,40.3,18,195,3250,female,2007
Adelie,Torgersen,NA,NA,NA,NA,NA,2007
The fourth record is useful: it still has a species, island, and year despite missing measurements. Do not delete it simply to make the table look complete.
Keep the original file unchanged and prepare a separate publication copy. Use the field definitions to check the imported columns:
| Columns | Meaning and intended type |
|---|---|
species, island |
Category labels |
bill_length_mm, bill_depth_mm |
Numeric bill measurements in millimeters |
flipper_length_mm |
Integer flipper length in millimeters |
body_mass_g |
Integer body mass in grams |
sex |
Female or male when recorded |
year |
Integer study year: 2007, 2008, or 2009 |
Prepare a publication copy readers can interpret
Make the missing-value rule explicit. The source CSV uses NA, and the project’s data preview shows those entries as missing values. For a publication copy, you can convert fields whose entire value is NA to empty fields, then explain that blanks mean missing—not zero. Record the conversion in your methods note. Do not assume a publishing tool will recognize NA automatically; check the result. See empty fields, zero, and missing values for the distinction.
Keep units attached to the fields. Retain the _mm and _g headers. Leave values numeric: 3750, not 3,750 g. Put human-readable labels and explanations in accompanying documentation.
Check the full file before sharing. With no rows removed, the expected total is 344 data rows, excluding the header. The project’s species summary provides a second check:
- Adelie: 152 rows
- Chinstrap: 68 rows
- Gentoo: 124 rows
Reopen the exported publication copy and verify these counts, all eight columns, and the fourth record’s missing fields.
Build a short dataset reference sheet containing the source link, CC0 statement, download date, study period, field definitions, and your missing-value conversion. State whether you published all records or a subset.
Publish, then check the reader’s view
TablePage supports CSV uploads and creates a public dataset page with a filterable table. Upload only the reviewed publication copy. For any other dataset, check for sensitive information and redistribution rights before making it public.
Use a descriptive title such as “Palmer Penguins: Measurements from 2007–2009.” Keep the reference notes accessible alongside the table, whether on the dataset page or the webpage where you share it.
Before distributing the link, check that:
- All 344 records remain available when filters are cleared.
- Filtering by species produces the expected counts.
- Numeric fields behave as numbers, not text categories.
- Missing measurements remain missing rather than becoming zeros.
- Any chart states its units and makes exclusions of missing values clear.
Frame the page as measurements from a particular study—not a census of Antarctic penguins. If you summarize body mass by species, report the number of non-missing measurements used alongside each result. That gives readers both an explorable table and enough context to check its interpretation.