Probably Pancakes
Description
An app that takes a user input recipe and classifies it as pancakes or not pancakes using a Random Forest algorithm trained on a data set of 3,200 recipes I collected from the web in 2018.
Goal
My household has a recipe written on a scrap of paper labeled “probably pancakes” (spoiler alert: it’s definitely pancakes). I have a weakness for over-engineered solutions to silly problems, so I set out to create a highly polished app to identify pancake recipes.
Outcome
I submitted the app to the 2nd Annual Shiny competition hosted by RStudio. Though the app did not win, it was invited to be included in the RStudio Shiny gallery (forthcoming).
Notable Features
- Flexible and dynamic UI driven by Shiny modules. User can input any number of desired ingredients, while the code remains succinct and easily maintained on the back end.
- Custom CSS, demonstrating the level of polish possible for client-facing products built with Shiny.
Technology Used
- Shiny
- Shiny modules
- data.table
- HTML/CSS