This repo contains my #rstats, data science & stats illustrations shared on my twitter account (@allison_horst).
All of this artwork is 100% available (and encouraged!) for open use by CC-BY license. That means: Hooray! I'm so happy that you want to share this artwork - especially if it helps when teaching R/rstats/stats. For most of the artwork you can just cite with "Artwork by @allison_horst". Please note some series where a different citation is requested (e.g. for collaborative projects).
This work is licensed under a Creative Commons Attribution 4.0 International License.
This artwork is available for free to anyone who wants to use it for your teaching, learning, presentations, and more. If you are a teacher and feel that your course benefits from the artwork, and you can do so without stress or burden, please consider a donation to Indigenous Women Hike. There are a number of ways to donate to Indigenous Women Hike, see options (including PayPal & Venmo) here: https://linktr.ee/iwh.
I am an Assistant Teaching Professor at the Bren School of Environmental Science and Management and it would be wonderful to include information about where and how this artwork is being used for my reviews. I would be so grateful if you could complete this very short (5 questions) Google Form, this will really help me out!
Questionnaire: Allison Horst R/stats artwork use
Thank you to Rob Hyndman for feedback & suggestions on this series!
Tidy Data series (a collaboration for the Openscapes blog with Dr. Julia Lowndes!)
Please cite the following Tidy Data illustrations as: "Illustrations from the Openscapes blog Tidy Data for reproducibility, efficiency, and collaboration by Julia Lowndes and Allison Horst"
- What is tidy data?
- Like families...
- Comparison of work benches:
- Friends with similar tools:
- Easier for automation & iteration!
- And it makes all other tidy datasets seem more welcoming!
- So make friends with tidy data!
R Knowledge Rollercoaster:
Derivatives thread:
usethis (seriously...):
Faces of debugging:
Monster supporters:
beepr let's you pick and play a notification sound when your code/analysis is done running:
beepr blank:
broom makes messy model / statistical outputs into tidy tibbles:
broom blank:
dplyr::mutate creates or transforms a variable (column) while keeping the existing ones:
mutate blank:
dplyr: get your data wrangling on.
dplyr::across() makes it easy to apply a function (or functions) across selected columns!
across blank:
dplyr::case_when() for friendly if_else statements:
dplyr::filter() to subset rows based on your conditions:
filter blank:
dplyr::relocate: a friendly function for moving columns around (in dplyr 1.0.0)!
relocate blank:
gganimate: get a little action in(to your graphs)...
gganimate blank:
ggplot2 for visual data exploration:
ggplot2 exploratory blank:
...and use ggplot2 for creating beautiful data masterpieces!
ggplot2 masterpiece blank:
here for more peaceful (file) paths:
here blank:
The janitor package contains multiple user-friendly functions for cleaning messy data, including clean_names() to update all of your column names to a nice case of your choosing (snake_case! lowerCamel! UpperCamel! SCREAMING_SNAKE! ...and more) all at once:
Use lubridate to work more easily & intuitively with dates & times:
lubridate time control blank:
Like lubridate_ymd() to easily parse year/month/day data!
lubridate ymd blank:
Use readr::parse_number() to just keep the numeric parts, & remove characters:
parse_number blank:
Part of tidymodels, the parsnip package creates standardized syntax across model engines:
parsnip blank:
Easily arrange and combine ggplots with patchwork!
patchwork blank:
You can do it!
R first-then blank:
Use @tylermorganwall's rayshader package to create amazing 3D maps and graphs!
Use recipes to streamline data preprocessing for stats & machine learning models:
Create reproducible examples to get (and give) help more easily with reprex!
reprex blank:
Get your code, text & outputs in the same (reproducible) place with Rmarkdown:
rmarkdown rockstars blank:
Be an Rmarkdown knitting wizard.
rmarkdown wizards blank:
Do your data sci like it's going to need an alibi with Rmarkdown:
reproducibility court blank:
Use the sf package for simpler spatial data analysis with geometries that stick to attributes:
sf blank:
Soon to be pivot_wider() & pivot_longer()! tidyr::spread() & gather():
stringr::str_squish() removes whitespace before and after strings, and reduced repeated interior whitespace to a single space (see also: str_trim()):
str_squish blank:
Blast off into the...
For #rstats and friends!
rstats blank:
Thanks, #rstats community!
codehero blank:
If you bring group_by() to the party, don't forget dplyr::ungroup()
ungroup blank:
The following illustrations are in Hadley's ACM talk, which you can watch HERE. Please cite the following artwork with "Illustrations from Hadley Wickham's talk "The Joy of Functional Programming (for Data Science)."
For looped:
Wrangler:
purrr feels like:
Presenting results:
I'm building this library of samples, faces & arms so that statistics teachers can create their own fun, charismatic samples to include in stats lectures, slides & materials. The files below contain different graphs (dotplots, histograms, more to come) with matching arms doing different things, along with a file of faces you can add on top to give them some personality. I recommend playing with transparency, brightness, cropping & size in whatever program you use to piece these together! Working on making these PNGs & SVGs.
Choose the expression to add to your sample:
More coming, feel free to send suggestions.
I start with "pretend you are this whale shark..."
For the love of pie charts:
Creatures and their distance matrix:
Find the clusters with the minimum distance between elements in them & merge:
Repeat!
Ta-da!
Meet your MLR teaching assistants:
Interpret coefficients for categorical predictor variables:
And for continuous predictor variables:
Or make predictions using the regression model:
Understand residuals:
And check for residuals normality:
in_case_you_forget:
Release the disco data:
Type I errors:
Type II errors:
Normality?
Continuous & discrete data:
Nominal, ordinal & binary data:
The expanded version of the classic Grolemund & Wickham R4DS workflow, including environmental data & sci comm bookends! Envisioned by Dr. Julia Lowndes for her useR!2019 keynote. Please cite this illustration as: "Updated from Grolemund & Wickham's classis R4DS schematic, envisioned by Dr. Julia Lowndes for her 2019 useR! keynote talk and illustrated by Allison Horst."
Dog & whale training art:
Make a data shark:
Data to make the shark is HERE. Created with drawdata.xyz.
-
Español: rstats-artwork-ES
-
Brazilian Portuguese: rstats-artwork-PT
-
Please submit translations as an issue!
Thank you to all the R developers, maintainers, contributors, teachers and communicators who actually MAKE all of these amazing packages and documentation that have inspired this #rstats artwork. When I create an illustration with your package it's with immense gratitude for how your hard work has allowed me to do mine (using and teaching #rstats) more efficiently, more clearly, more reproducibly....just plain better. THANK YOU!