Tidyfun: A New Framework for Representing and Working with Function-Valued Data

• Activity Number: 59 - Invited E-Poster Session I
• Type: Invited
• Date/Time: Sunday, August 8, 2021 : 5:45 PM to 6:30 PM
• Sponsor: Biometrics Section
• Author(s): Jeff Goldsmith* and Julia Wrobel and Fabian Scheipl
• Companies: Columbia University, Department of Biostatistics and University of Colorado Denver, Department of Biostatistics and Ludwig Maximilians University Munich
• Keywords: Functional data; software; tidyverse

Abstract

A new R package, tidyfun, provides accessible and well-documented software that makes functional data analysis in R easy – specifically data wrangling and exploratory analysis. This is achieved by introducing a new data type. Vectors of this type can be operated on using many standard functions (+, mean, etc.) as well as several new functions in tidyfun (tf_smooth, tf_where). Crucially, vectors of this class can be included in data frames containing both scalar and functional data, enabling data manipulation and visualization using tidyverse tools. This approach is connected to the conceptual framework in functional data analysis, which assumes that complete functions are the unit of observation; with tidyfun, full curves sit alongside numeric, factor, and other observations on the same subject. We discuss the available feature set as well as forthcoming extensions and show some application examples.