Probably Pancakes
Authors: Matthew Rogers
Working with Shiny more than 1 year
Abstract: An app for predicting whether or not an input recipe is or is not pancakes.
Full Description: In our recipe book, my wife and I have a recipe labeled “probably pancakes”. My wife transcribed the recipe at some point in college without a label, and eventually it acquired its probabilistic identification. I suppose some people would be satisfied that since the recipe yields fluffy, flat cakes that go really well with maple syrup it is, in fact, a pancake recipe.
I'm not one of those people. I have a deep affinity for thoroughly engineered solutions to trivial problems, and to that end I created a data set suitable for machine learning from recipes collected from allrecipes.com and trained a model to predict whether a given recipe was or was not pancakes.
The app building process posed two primary challenges. The first was building an interface that would accept an arbitrary number of inputs determined by the user. The second, self-imposed, was a desire to have a highly customized and polished UI. The first was achieved through the use of modules and lapply, the second with the help of the excellent w3schools.com.
I provided the recipe 'probably pancakes' as the default inputs in the app, but you can input a custom recipe or an allrecipes.com recipe URL in case you too have a burning need to know whether or not you are, in fact, making pancakes.
Category: Other
Keywords: machine learning, fun
Shiny app: https://matthew-rogers.shinyapps.io/ProbablyPancakes/
Repo: GitHub - matthewjrogers/pancake_prediction_app: Companion to web scraping and ML pancake project
RStudio Cloud: Posit Cloud
Thumbnail:
Full image:
