Analysis of Participant-Level Characteristics Predicting Adherence to Long-Term EMA and Fitbit Monitoring

This directory contains R scripts, data files, and analysis reports relatedto the analysis of adherence. FILE OVERVIEW DATA FILES* data.xlsx Main subject-level dataset. * data-fitbit.xlsx Dataset with detailed Fitbit-derived measures. DATA LOADING AND PREPROCESSING* data.R R script for loading and preprocessing the main dataset. * data-fitbit.R R script for loading and preprocessing the detailed Fitbit dataset. DESCRIPTIVE AND BASELINE ANALYSES* summary.Rmd R Markdown document providing a basic description of the datasets. * baseline.Rmd R Markdown document with baseline characteristics and missing-value analysis, including evaluation and imputation of missing values in SWL (new variable SWLlm.predicted). * baseline-by-burst.Rmd R Markdown document with baseline characteristics stratified by burst number. MODELING AND VARIABLE IMPORTANCE* importance.Rmd R Markdown analysis template for evaluating variable importance using stepwise regression and random forest models. PROJECT CONFIGURATION AND BUILD SYSTEM* Makefile.R Project definition for the rmake package; generates the GNU Makefile. * Makefile File dependencies and compilation commands used by GNU Make to generate all project results. * Rproject.Rproj RStudio project file. ------------------------------------------------------------------------------REQUIREMENTS------------------------------------------------------------------------------ 1) INSTALL REQUIRED R PACKAGES Run the following commands in R: install.packages(c( "tidyverse", "knitr", "rmake", "caret", "randomForest", "fastDummies", "broom", "devtools" )) devtools::install_github("beerda/hammer") devtools::install_github("beerda/mbrtools") 2) GENERATE PROJECT ANALYSES Run the analysis pipeline using rmake: make() This command generates all results defined in the project Makefile.------------------------------------------------------------------------------