wtf is cocktails-demo?

The cocktails-demo project is a small web app that helps you pick a cocktail to make. You open it in your browser, see three random drink ideas, and can narrow things down by choosing a base spirit (like gin, vodka, or rum) and the kind of flavors you're in the mood for (bitter, fruity, refreshing, smoky, and so on). Each cocktail comes with a short description and the ingredients you'd need. Under the hood, the app is built with R using a framework called Shiny, which lets you turn R code into an interactive web page without writing separate frontend code. The cocktail data lives in a structured text file format called YAML, where each drink has fields for its name, description, primary spirit, ingredient list, and flavor tags. The app reads that file, displays three cocktails at random, and re-filters the list whenever you pick a spirit or flavor. It's all one self-contained file, which keeps things simple. The audience here is really anyone who wants cocktail inspiration and happens to have R installed, a home mixologist, a data science instructor demoing Shiny, or someone curious about how AI tools can build small apps. It's less a production product and more a teaching example or personal utility. What's notable is that the app was largely generated by ChatGPT, with light manual edits for correctness and appearance. The README includes the actual prompts used to generate both the cocktail data and the app itself, making it a useful reference if you want to see how to describe an app idea clearly enough for an AI to build it. The project trades polish and scale for simplicity and transparency about its origin.