rstudio::conf(2020L) #Learning R with humorous side projects

I’m incredibly thankful for the opportunity to speak at rstudio::conf 2020 about all of the silly projects I have created to help me learn R. It’s not often that I get an opportunity to speak about dinosaurs, drinking games, LEGO bricks, or the Golden Girls in a professional setting, let alone all 4 in the same talk!

Check out the video of the talk here 1 and the pdf of the slides can be found here.

The slides, however, are just a preview of those projects. Below, I aggregate the original posts, tweets, and repositories from the projects featured during the talk.

The Golden Girls drinking game with {tidytext}

Optimizing the Golden Girls drinking game with {tidytext} is my OG project for learning valuable skills with a ridiculous topic.

  • Original blog post has details about my approaches to this project. (Note: there’s some colorful language in the drinking game rules that I excluded from my conference talk.)

  • Check out Julia Silge & David Robinson’s book Text Mining with R that has all the information you’ll need to get started with the package.

My other drinking games don’t have any official blog posts, but I use the same method using scripts and data from The Good Place, The Office, and Jurassic Park.

Jurassic Park character paths with {gganimate}

  • Short blog post with an overview of this project and additional charts of the Jurassic Park character paths.

  • Github repo for Thomas Lin Pedersen’s {gganimate} package should have everything you need to get started with the package. In most cases, you’ll just add one line of script to your ggplots.

  • I also learned about {gganimate} with a small art project to convert images into spiral line drawings. Check out this blog post for more details and code snippets.

Datasaurs Twitter bot

Datasaurs is a Twitter bot that takes a time series of mortality data and finds a dinosaur (or other animal) outline that is correlated with it, then redraws the dinosaur using the data series. The bot is currently offline as I work on version 2 and get a new server up.

  • Twitter home of Datasaurs and short blog post explaining the concept.

  • GitHub repo for Datasaurs which should be reproducible for you (after some work) if you grab some images from PhyloPic and have an API key. My main goal for version 2 will be to wrap the correlations & drawing into a package, so anyone can find these silly correlations.

  • This project used so many packages to get to work. Each one of them has been very valuable in my career as a data scientist.

Generating new dinosaur names with {keras}

I learned how to start with {keras} to do deep learning by using recurrent neural networks to generate new dinosaur names from a list of a few thousand extinct animal names I scraped from Wikipedia.

  • Blog post explaining my process and my learnings. To make the output more fun, I used {ggplot2} and images from PhyloPic to display the names on fake phylogenetic trees.

  • This project was a literal clone of Jacqueline Nolis’ project to generate offensive license plates with RNN. She explains it way better than I do.

  • Jacqueline also has a book coming out soon, written with Emily Robinson! Check it out: Build a Career in Data Science.

  • A follow-up blog post by me, using {keras} to distinguish between dinosaurs names and Pokemon names.

LEGO bricks in R with {brickr}

{brickr} is my R package, where I try to join the R nerd and LEGO nerd communities in one place. The package can be used to generate LEGO mosaics from images, build 3D LEGO brick models from coordinates, or add a LEGO twist to your ggplots.

I plan on working on this package heavily over the next few months. Let me know if you want to help!

  • brickr.org has everything to you need to get started with the project. It’s not yet on CRAN, so download using {remotes}.
#install.packages("remotes")

remotes::install_github("ryantimpe/brickr")

Thank you!!!


  1. For the full experience, play the drinking game rules for my talk while watching!
Avatar
Ryan Timpe
Data Science | Economics

Related