Example datasets

← Contents

The manual uses the public datasets in the repository's demo/examples/ for its basic operations and charts. Download them from Datasets, sources and licenses. The first row holds column names. Open compressed files directly, without extracting them.

File	Rows	Contents and uses
`iris.csv`	150	Three iris species, 50 flowers each. Basic operations, correlation and grouping. Lengths and widths are in cm.
`penguins.csv`	344	Three penguin species. Box plots of body mass and missing values. Lengths are in mm; mass is in g.
`hour.csv.xz`	17,379	Hourly Bike Sharing records. Date histograms and density of temperature against rentals.
`flights.csv.xz`	336,776	Flights departing New York in 2013. Departure and arrival delays in a larger table. Delays are in minutes; negative values mean early departures or arrivals.

Try Iris

Run these commands from the repository root. In the installed GUI use File → Open..., or Open... in the web app, to select iris.csv.

tableskim demo/examples/iris.csv
tableskim -t demo/examples/iris.csv
tableskim --summary demo/examples/iris.csv

The columns are sepal_length, sepal_width, petal_length, petal_width (all in cm), and species.

Search for setosa: there are 50 hits.
Apply Filter conditions species = virginica and petal_length ≥ 6 with Match: All: 11 of 150 rows remain. Use Clear Filter to reset.
Open a histogram of petal_length and choose Stack by → species to see each species' contribution.
In a scatter plot choose X → petal_length, Y → petal_width, Color by → species. The overall correlation is about 0.9629. It also reflects differences between species; filter to one species to study its own relationship.
In Group By, group by species, aggregate petal_length, and select mean, min, max. The means are 1.462 (setosa), 4.260 (versicolor), and 5.552 cm (virginica).
For SQL, use Convert to Parquet / DuckDB... (P in the TUI) and open iris.parquet. The query in Running SQL gives the same means.

See GUI, TUI or Web for the controls, and Charts and Group By for how to read the results.

Check missing values and units

penguins.csv has two missing body masses and two missing flipper lengths. A plot of flipper_length_mm against body_mass_g draws 342 penguins and skips two rows. Overall correlation includes size differences between species; distinguish it from correlation within a species.

In hour.csv.xz, cnt counts rentals in one hour and temp is normalized temperature, not degrees Celsius. season and weathersit are numeric category codes. Box plots need text on X; do not interpret those codes as continuous measurements. dteday is a date, with multiple hourly records for the same day.

flights.csv.xz represents missing values with the text NA. For delay analysis, filter with dep_delay is a number and arr_delay is a number to keep numeric pairs.

Examples specific to file formats, such as Excel date formats and multiple sheets, also use generated files in demo/data/. The public dataset records remain unchanged; screenshots show views and summaries. Iris and Bike Sharing are CC BY 4.0; Penguins and Flights are CC0 1.0. The link above includes sources and credits.

Iris: Fisher, R. (1936), DOI: 10.24432/C56C76. Bike Sharing: Fanaee-T, H. (2013), DOI: 10.24432/C5W894. Both are CC BY 4.0. Penguins: Kristen B. Gorman / Palmer Station Antarctica LTER, palmerpenguins (Horst, Hill and Gorman, 2020). Flights: Hadley Wickham, nycflights13 (records from the US Bureau of Transportation Statistics). Both are CC0 1.0.