Description:
This talk takes a deep dive into the world of testing shiny applications, covering strategies for determining what needs to be tested to ensure your app's functional accuracy and reliability. Shiny's reactive programming model requires specialized approaches to testing, distinct from traditional R package testing.
I'll begin by describing the application of unit, integration, and end-to-end testing in Shiny apps. Then we'll move into the role of testing in development and highlight the need for code organization when designing your testing suite.
The focus then shifts to practical aspects of testing by illustrating tools like test data and fixtures. Attendees will see example tests for utility functions, modules, and tests to capture specific user interactions (and confirm the application behaves as expected).
Finally, I will demonstrate advanced topics like test coverage, what it measures, and why 100% coverage isn't necessarily a metric worth chasing.
This presentation will give R developers a toolkit of testing techniques (building a test suite, writing effective tests, and automated testing) regardless of their familiarity with Shiny.
Bio:
Martin Frigaard is a Data Visualization Engineer at Atorus, where he builds dashboards and tools for making data-informed decisions. Previously, Martin built statistical tools and dashboards at BioMarin, The Diabetes Technology Society, and other volunteer and non-profit organizations. He was one of the original contributing authors for Data Journalism in R, an award-winning resource hosted by Northeastern University's School of Journalism. Martin also previously taught data journalism at California State University, Chico. Before obtaining graduate degrees in Clinical Research and Exercise Science, Martin served as a US Army medic in Operation Enduring Freedom.
