Teal Builder - The Fastest Way to Build Reproducible Teal Apps

Abstract

This session will introduce Teal Builder, a powerful Shiny application designed to simplify the creation of Teal apps without requiring users to write any R code. The presentation will cover the purpose and functionality of Teal Builder, demonstrate how to run the app locally, and provide a live demonstration of its capabilities. Teal Builder not only streamlines the process of building Teal apps but also provides a platform for generalizing certain internal statistical methods and processes into reusable modules to be used by more studies. Teal Builder offers a user-friendly interface that guides users through data selection from multiple sources, module selection and compatibility validation based on the selected data. The app generates reproducible R code for running the Teal app or deploying it to Posit Connect, making the process efficient and accessible. It fulfills the unmet need of having an interactive data display for all studies, promoting Teal and R/Shiny adoption and usage to new R and non-R users. Teal Builder comprises three R packages teal.builder, teal.builder.modules, and teal.builder.checks. The teal.builder package contains functions for running the Shiny app and serves as the entry point for most users. The teal.builder.modules package offers a collection of predefined module specifications, including module names, descriptions, data requirements, and reproducible code. The teal.builder.checks package provides functions to ensure user data adheres to specifications, offering detailed feedback when necessary. During the demonstration, we will walk through the user journey of building a Teal app with Teal Builder. We’ll showcase how a complex app for exploring ADaM datasets, which typically requires around 500 lines of code, can be generated in just a few minutes. This session will highlight the efficiency and user-friendliness of Teal Builder, making it an invaluable tool for those working with Teal apps.

Type
Publication
Presented at 2024 Conference

Related