Problem: Needed a simple, fun app to demonstrate the capabilities of R Shiny as well as the concept of machine prediction for a Technology and Innovation Fair
Solution: Built a simple R Shiny app that allows players to play against a computer agent that uses experience from past rounds to predict what the players next move is.
Results:
  • Used by more than 50 players during Technology Fair
  • Prediction algorithm shows reasonable accuracy though I have not formally evaluated it.
Status: Live app available on shinyapps.io; Code available on Github
Skills Used: web development; machine learning; Google Sheets
Languages Used: R; HTML; CSS

I am a co-organizer of the BLS Data Science User Group. This year we sought to advertise our existence by setting up a booth at the BLS Technology and Innovation Fair. In addition to practical, BLS related data science projects, we wanted to have at least one simple, fun app that could be used to demonstrate the capabilities of R and R Shiny while also giving visitors to the booth an example of machine learning.  So over a weekend, I built this Rock, Paper, Scissors application.  I had to extend myself a bit on creating this because shinyapps.io does not allow persistent storage so in order to save previous player interactions, I had to learn how to store data in Google Sheets.  It’s that writing to Google Sheets that causes some performance issues with the app (e.g., images not displaying during each round).  It runs much better locally.

Leave a Reply

Your email address will not be published.