Intraindividual Dynamic Network of Affects

This repository hosts R code for a Shiny application developed as part of DynaMORE projct. It provides tools to fit an Exogenous Linear Autoregressive Mixed Effects model, LARMEx, to ecological momentary assessments (EMA).

How to use the app?

I. The online version

• larmexShiny

II. Running on local computer

• install R from r-project
• install RStudio from rstudio-desktop

II.I. Install as an R package

• to install packages from GitHub one needs the R package "remotes"
  • install.packages("remotes")
• install "larmexShiny""
  • remotes::install_github(repo = "spooseh/larmexShiny", ref = "master", dependencies = TRUE)
• load the package
  • library(larmexShiny)
• run in R command line
  • runLARMEx()

II.II. Download the source code and run locally

• two ways to get the codes:

  • i. through terminal by cloning the repository
    • git clone -b clone_and_run --single-branch https://github.com/spooseh/larmexShiny.git
  • ii. through a browser by downloading the repository
    • go to github.com/spooseh/larmexShiny
    • select the branch "clone_and_run"
    • download the code as a zip file ("larmexShiny-clone_and_run.zip")

• navigate to the local directory, larmexShiny, in RStudio

• optional: for a smooth work flow, make this folder, by setwd(), the working directory

• open packInstaller.R and press Source in RStudio

  • it runs the command source("packInstaller.R")
  • one could install the missing packages manually
  • see sessionInfo.txt for version information

• open the runLARMEx.R file and click the Run App at the top of your editor's menu

What you see

• an Rstudio window opens and shows the user interface
• you can click Open in Browser or enter the address next to it, http://127.0.0.1:6967, on a browser of your choice (6967 may be different in your case)

Instructions

• detailed instructions in larmexShiny/www/Instructions.md

Not interested in a GUI?

The following steps walk you through the workflow of fitting some simulated data and hopefully demonsterate how to treat your own. Once "larmexShiny" is installed in R:

• access the simulated data included in the package
  • simMood <- system.file("extdata", "simMood.csv", package="larmexShiny")
• create an object holding the data and settings
  • obj1 <- LARMExFit$new()
  • ?LARMExFit for more information
• assign raw data, long, single or multiple subjects, dataframe
  • obj1$rawData <- read.csv(simMood)
• variable name, subject ID, character
  • obj1$sjID <- "sjID"
• variable name, number of beeps, character
  • obj1$nDay <- "nDay"
• variable name, number of days, character
  • obj1$nBeep <- "nBeep"
• variable names, interacting moods (AR nodes), character
  • obj1$arList <- c("M1", "M2")
• variable names, external factors acting on meeds (EX nodes), character
  • obj1$exList <- "E"
• where to save results, default to HOME
  • obj1$savePath <- normalizePath("~")
  • toDir = file.path(obj1$savePath, "fitRes")
  • dir.create(toDir, recursive=TRUE, showWarnings=FALSE)
• extract data for a single subject
  • sj <- 1001
  • selCols <- c(obj1$nDay, obj1$nBeep, obj1$arList, obj1$exList)
  • sjData <- obj1$rawData[obj1$rawData[obj1$sjID]==sj, selCols]
• transform the data into a special format suitable for LARMEx
  • df2Fit <- prepData2Fit(obj1, sjData)
• set the formula using the transformed data (the column names are used)
  • setFormulaDf(obj1, df2Fit)
• fit the data and save the results
  • fitLmer1(obj1, sj, toDir)

To do ...

• exception handling
• more documentation

---