Data Sets from the History of Statistics and Data Visualization
Version 1.1.1 (2026-10-05)
The HistData package provides a collection of small data sets that are
interesting and important in the history of statistics and data
visualization. The goal of the package is to make these available, both
for instructional use (as examples, problem sets or projects) and for
historical research (extending or criticizing a previous analysis). Some
of these present interesting challenges, or opportunities to “show off”,
with graphics or analysis in R.
Many of the data sets contained here have examples which reproduce an historical graph or analysis. These are meant mainly as simple starters for more extensive re-analysis or graphical elaboration. If you are interested in any of these problems or data sets, there is lots of room to do better!
Some of these been featured in social media calls for participation, such as the 30 Day Chart Challenge. A new vignette describes this more generally as a HistData Challenge.
This is part of a program of research called statistical historiography (Friendly, 2007; Friendly & Denis, 2001; Friendly et-al, 2016) meaning the use of statistical methods to study problems and questions in the history of statistics and graphics. A main aspect of this is the increased understanding of historical problems in science and data analysis trough the process of trying to reproduce a graph or analysis using modern methods. I call this “Re-visioning”, meaning to see again, hopefully in a new light.
They are also used in our book, (Friendly & Wainer, 2021), A History of Data Visualization & Graphic Communication, Harvard University Press, ISBN 9780674975231. See also the companion website for this book.
If you are looking more widely for datasets to use for examples, teaching or research, check out Vincent Arel-Bundock’s Rdatasets package, with over 2500 datasets from various R packages, with this list of Available datasets.
There is another R aspect of the HistData project that should be noted here:
A great deal of “data sciency” work was involved in constructing this
package. This may not be evident in what you see in the resulting
HistData package you see, but for those interested in this form of
History Data Mining, here are some of the aspects involved:
In some cases, data had to be extracted from dusty historical documents, using a variety of techniques (web scraping, OCR of PDF files followed by conversion to a data set). Each problem had its own toolbox, inside R or outside. In many cases, transcription errors had to be corrected with code or manually. This is much simpler today with many tools, and the possibility to enlist an AI companion. But, whatever you use, you’d better check the work. Some of the more interesting stories of statistical historiography involve uncovering transcription errors.
Digitization of data from an image. When the data is only available in an image, you can extract it with more modern tools than were available when I started this in 2007. Among the newer ones: WebPlotDigitizer, PlotDigitizer. But, where accuracy is important, you should think about the reliability and valitity of what you got from an image (Drevon etal. (2016)).
Conversion of text-based data sets to a CSV file and then to an
.RData file with proper column names. Ever you seen a Unix .shar
(shell archive) file? Well, I have. What about an old codebook
prepared for a survey? Modern R packages can read data in a much wider
variety of formats today, but older material still presents
challenges.
NameCleaning for variable names: Some of this can now be done
using e.g.,janitor::clean_names(), or, in some cases, manually
editing an excel file you got. What to do about
very_long_variable_names or those “abbrvd 2 shrt”? But you should
also adopt a consistent naming convention across your datasets in a
project: camelCase, snake_case, PascalCase, kebob-case;
hopefully never SCREAMING_SNAKE_CASE, unless for a constant like
MAX_ITER.
Some similar considerations apply to the values of character or
factor variables in a dataset, and also how these will be ordered in
tabular or graphic displays. As one example, levels of education, with
values "Low", "Medium", "High" will be sorted “H, L, M” by
default. You can recode or assign levels
c("1:Low", "2:Medium", "3:High") or create them as ordered()
factors.
Applying Type-conversion, e.g., chr to factor or ordered
required thinking about what a variable represented. In some cases,
appropriate contrasts for factors were constructed to facilitate
re-analysis. Pay attention to StringsAsFactors.
Tidying data.frames: Initially, most of this was done in base R,
but as they came along tidyverse tools made these tasks much easier,
e.g., conversion of long <–> wide, separating implicit columns, like
“1800-1850”, abbreviations of character string labels, …
Documentation: The thankless task? No – considerable effort was made to give detailed descriptions, notes on methods, executable examples, references to original sources and analyses, …
Dataset documentation: Originally, I wrote most of the
documentation for datasets using utils::prompt(dataset). There isn’t
anything like this to generate roxygen documentation for a dataset, so
I wrote something
use_data_doc.R
to do this.
Documentation style: In this release, all the documentation was
converted first from manually written .Rd files, to roxygen format
using Rd2roxygen.
This package does a reasonable job, but gets a few things wrong. E.g.,
one that required a complicated regex to convert
\item{list("VarName")} to the correct format
\item{\code{VarName}}. But documentation is even easier to edit and
maintain using markdown format, as converted again by
roxygen2md.
Provenance: Where possible, a dataset’s documentation should trace
a clear path from its original source to the version included here:
that’s the Historian’s Hat. Not just “digitized from X,” but how,
and what’s still unverified or questionable. Perozzo is a case in
point: its immediate source (a GitHub redraw of the stereogram) was
traced and confirmed byte-identical, but the actual digitization from
Perozzo’s 1880/1881 plate remains unknown. The current documentation
says so, rather than implying more confidence than the data warrants.
Ask AI: A few tasks were aided by Claude Sonnet 4.5 and bros.
I wanted to add @concept descriptors to datasets reflecting the
statistical and graphical concepts that each dataset can be used to
illustrate. It made a list, which I heavily edited, and also an R
script to add these to the
documentation.
In a couple of recent additions like
Federalist
I used Claude as a skillful research assistant. We worked together
to access the Gutenberg
text using the
gutenbergr
package, devise a schema for datasets that would be useful for
analysis or graphs, and then incorporate into the package. All this
recorded and described in the Federalist.R
script used to import the data and create
the R datasets.
Get the released version from CRAN
install.packages("HistData")
The development version can be installed to your R library directly from github or my R-universe via:
install.packages('HistData', repos = 'https://friendly.r-universe.dev')
remotes::install_github("friendly/HistData")
Here are the data sets in the package, with links to their documentation. Some topics are represented by two or more data sets.
# link dataset to pkgdown doc
refurl <- "https://friendly.github.io/HistData/reference/"
dsets <- vcdExtra::datasets("HistData") |>
dplyr::select(Item, Title) |>
dplyr::mutate(Item = glue::glue("[{Item}]({refurl}{Item}.html)"))
library(tinytable)
# tt(dsets) |>
# format_tt(j = 1, markdown = TRUE) |>
# style_tt(j = 1, bootstrap_css = "width: 30%;") |>
# style_tt(j = 2, bootstrap_css = "width: 70%;")
tt(dsets, width = c(.2, .8)) |>
format_tt(j = 1, markdown = TRUE)
| Item | Title |
|---|---|
| Arbuthnot | Arbuthnot’s Data on Male and Female Birth Ratios |
| Armada | La Felicisima Armada |
| Bowley | Bowley’s data on values of British and Irish trade, 1855-1899 |
| Breslau | Halley’s Breslau Life Table |
| Cavendish | Cavendish’s Determinations of the Density of the Earth |
| ChestSizes | Chest measurements of Scottish Militiamen |
| ChestStigler | Chest measurements of Scottish Militiamen |
| Cholera | William Farr’s Data on Cholera in London, 1849 |
| CholeraDeaths1849 | Daily Deaths from Cholera and Diarrhaea in England, 1849 |
| CushnyPeebles | Cushny-Peebles Data: Soporific Effects of Scopolamine Derivatives |
| CushnyPeeblesN | Cushny-Peebles Data: Soporific Effects of Scopolamine Derivatives |
| Dactyl | Edgeworth’s counts of dactyls in Virgil’s Aeneid |
| DrinksWages | Elderton and Pearson’s (1910) data on drinking and wages |
| Ebbinghaus | Ebbinghaus’ Forgetting Curve and Three Replications |
| EdgeworthDeaths | Edgeworth’s Data on Death Rates in British Counties |
| Federalist_corpus | The Federalist Papers: corpus metadata, marker-word counts, and text |
| Federalist_markers | The Federalist Papers: corpus metadata, marker-word counts, and text |
| Federalist_text | The Federalist Papers: corpus metadata, marker-word counts, and text |
| Fingerprints | Waite’s data on Patterns in Fingerprints |
| Galton | Galton’s data on the heights of parents and their children |
| GaltonFamilies | Galton’s data on the heights of parents and their children, by child |
| Guerry | Data from A.-M. Guerry, “Essay on the Moral Statistics of France” |
| HalleyLifeTable | Halley’s Life Table |
| Handwashing.clinics | Semmelweis’ Data on Handwashing and Deaths from Childbed Fever |
| Handwashing.hospital | Semmelweis’ Data on Handwashing and Deaths from Childbed Fever |
| Handwashing.monthly | Semmelweis’ Data on Handwashing and Deaths from Childbed Fever |
| Jevons | W. Stanley Jevons’ data on Numerical Discrimination |
| Langren.all | van Langren’s Data on Longitude Distance between Toledo and Rome |
| Langren1644 | van Langren’s Data on Longitude Distance between Toledo and Rome |
| Macdonell | Macdonell’s Data on Height and Finger Length of Criminals, used by Gosset (1908) |
| MacdonellDF | Macdonell’s Data on Height and Finger Length of Criminals, used by Gosset (1908) |
| Mayer | Mayer’s Data on the Libration of the Moon. |
| Michelson | Michelson’s Determinations of the Velocity of Light |
| MichelsonSets | Michelson’s Determinations of the Velocity of Light |
| Minard.cities | Data from Minard’s famous graphic map of Napoleon’s march on Moscow |
| Minard.temp | Data from Minard’s famous graphic map of Napoleon’s march on Moscow |
| Minard.troops | Data from Minard’s famous graphic map of Napoleon’s march on Moscow |
| Nightingale | Florence Nightingale’s data on deaths in the Crimean War |
| OldMaps | Latitudes and Longitudes of 39 Points in 11 Old Maps |
| PearsonLee | Pearson and Lee’s data on the Heights of Parents and Children by Gender |
| Perozzo | Perozzo’s Swedish Population Survivorship Data |
| Playfair1824 | Playfair’s Linear Chronology |
| PolioTrials | Polio Field Trials Data |
| Pollen | Pollen Data Challenge |
| Prostitutes | Parent-Duchatelet’s time-series data on the number of prostitutes in Paris |
| Pyx | Trial of the Pyx |
| Quarrels | Statistics of Deadly Quarrels |
| Saturn | Laplace’s Saturn data. |
| Snow.dates | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Snow.deaths | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Snow.deaths2 | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Snow.polygons | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Snow.pumps | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Snow.streets | John Snow’s Map and Data on the 1854 London Cholera Outbreak |
| Virginis | John F. W. Herschel’s Data on the Orbit of the Twin Stars gamma Virginis |
| Virginis.interp | John F. W. Herschel’s Data on the Orbit of the Twin Stars gamma Virginis |
| Wheat | Playfair’s Data on Wages and the Price of Wheat |
| Wheat.monarchs | Playfair’s Data on Wages and the Price of Wheat |
| Yeast | Student’s (1906) Yeast Cell Counts |
| YeastD.mat | Student’s (1906) Yeast Cell Counts |
| ZeaMays | Darwin’s Heights of Cross- and Self-fertilized Zea May Pairs |
The Horsekicks package contains the classic data from von Bortkeiwicz, “Death by the kick of a horse in the Prussian Army”, with additional data on deaths by falling from a horse and by drowning.
Rdatasets is a collection of over 2500 datasets culled from CRAN packages to make them broadly accessible for teaching and statistical software development.
The lattice package contains
the Minnesota barley data used by Cleveland (1993) in developing
Trellis graphics and ideas behind conditioning (faceted) plots.
Wright (2013) reanalysed this and extended the datasets in the
agridat package, which also
contains a wide variety other datasets from agricultural experiments,
some of historical interest.
Contributions of new data sets, corrections and examples are welcome. See the contributing guide for what fits, how to document a data set’s source, and the rules on copyright and permissions.
Please note that the HistData project is released with a Contributor
Code of
Conduct. By
contributing to this project, you agree to abide by its terms.
Over the years, many people have contributed new data sets, offered corrections, suggestions, or documentation examples. They are appreciatively listed below:
David Bellhouse, Brian Clair, Stephane Dray, Luiz Droubi, Antoine de Falguerolles, Monique Graf, James Hanley, Ivan Lokhovm Peter Li, Dennis Murphy, Jim Oeppen, James Riley, John Russell, Neville Verlander, Hadley Wickham.
Cleveland, William S. (1993) Visualizing Data. Hobart Press, Summit, New Jersey.
Drevon, D., Fursa, S. R., & Malcolm, A. L. (2016). Intercoder Reliability and Validity of WebPlotDigitizer in Extracting Graphed Data. Behavior Modification, 41(2), 323–339. https://doi.org/10.1177/0145445516673998
Friendly, M. (2007). A Brief History of Data Visualization. In Chen, C.,
Hardle, W. & Unwin, A. (eds.)
Handbook of Computational Statistics: Data Visualization,
Springer-Verlag, III, Ch. 1, 1-34.
Preprint
Friendly, M. & Denis, D. (2001). Milestones in the history of thematic cartography, statistical graphics, and data visualization. Web stite: http://datavis.ca/milestones/
Friendly, M. & Sigal, M. & Harnanansingh, D. (2016). “The Milestones Project: A Database for the History of Data Visualization,” In Kostelnick, C. & Kimball, M. (ed.), Visible Numbers: The History of Data Visualization, Ashgate Press, Chapter 10. Preprint
Friendly, M. & Wainer, H. (2021). A History of Data Visualization and Graphic Communication, Harvard University Press, ISBN 9780674975231. Companion web site
Wright, K. (2013). Revisiting Immer’s Barley Data. The American Statistician, 67(3), 129–133