In hamsamilton/cgms.analysis: Functions for CGMS analysis

Quick vignette to demonstrate cgms workflow

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Load in data. cgms contains raw data. census contains info on each subject and trial

library(dplyr)
# Data preprocessing #####

#Load in census
census <- rio::import("~/cgms_analysis/cleaned data/visitinfo.tsv")
#fix datatypes
cgms <- rio::import("~/cgms_analysis/cleaned data/cgms.csv")

Use 2cgmsdf to combine census and cgms together and tidy the data for analysis. This is the datafrane object other functions use.

cgms.df <- cgms.analysis::tocgmsdf(cgms,census)
print(head(cgms.df))

These cgms.df objects can and should be filtered to the subjects and runs that are relevant to your question. We use the nulldist function to add a nulldistribution to the dataframe. There are several to chose from depending on what You need your null distribution to be.

cgms.df.2124.V2PV8P <- cgms.df %>%
  filter(projno == 2124) %>%
  filter(visit %in% c("V2_P","V8_P")) %>%
  cgms.analysis::addnulldist()

From here, You can plot visualize all the runs for your cgms.df object, or calculate the sliding sample entropy using one of the entropy functions

cgms.analysis::plotruns(cgms.df.2124.V2PV8P,"/Users/samuelhamilton/cgms_analysis","name")

cgms.analysis::CalcEn(df = cgms.df.2124.V2PV8P,varlist = c("Group"))
cgms.analysis::CalcEnSlide(df = cgms.df.2124.V2PV8P,varlist = c("Group"))
head(cgms.analysis::CalcEnSlideDist(df = cgms.df.2124.V2PV8P,varlist = c("Group")))