EPGminer Example Workflow • epgminer

The following R code is illustrative of how to use this software on our data to reproduce calculations. In essence, data and annotations are read in using the read_epg data function. The label_ana function is used to combine the raw data and annotation files. Then wave_topfreq and wave_volts are used to calculate frequency and relative amplitude of volts respectively.

library(epgminer)

# The time needed to run this script is dependent on the number of datasets

# load necessary packages
library(tidyverse)
library(epgminer)

# Set working directory to folder containing dataset folders and ANA files
folder <- "insert_here"
setwd(folder)
# list out ANA files 
label <- list.files(pattern = ".ANA")

# Read all datasets into a list
raw_dat <- list()
lab_dat <- list()
for (i in 1:length(label)) {
  # find specified raw hourly files
  set <- str_sub(label[i], end = -5)
  
  # read in the specified dataset
  raw_dat[[i]] <- paste(set, "/", list.files(paste(set, "/", sep = "")), sep = "") %>%
    map_df(~epgminer::read_epg(.))
}

# for each dataset, combine raw and ANA
for (i in 1:length(label)) {
  ana <- epgminer::read_epg(label[i], extension = "ANA")
  # lab dat contains fully labeled tibble - time, volts, wavform
  # to be used for metrics
  lab_dat[[i]] <- epgminer::label_ana(raw_dat[[i]], ana)
  # track when data has been labeled 
  print(i)
}

# Now calculate top frequencies for each waveform
for (i in 1:length(lab_dat)) {
  set <- str_sub(label[i], end = -5)

  # returns tibble with waveform, frequency columns
  # pds (subform independent) are on bottom
  out <- epgminer::wave_topfreq(lab_dat[[i]])

  # save frequency results in a csv
  write.table(out, file = paste(set, "_frequencies.csv", sep = ""),
              row.names = FALSE, col.names = TRUE, sep = ",")
  # track table writing
  print(paste(set, "_frequencies.csv", sep = ""))
}

# voltage metrics ----
for (i in 1:length(lab_dat)) {
  set <- str_sub(label[i], end = -5)
  
  # returns tibble with waveform, duration columns
  # pds (subform independent) are on bottom
  out <- epgminer::wave_volts(lab_dat[[i]])
  
  # save results in a csv
  write.table(out, file = paste(set, "_volts.csv", sep = ""),
              row.names = FALSE, col.names = TRUE, sep = ",")
  # track table writing 
  print(paste(set, "_volts.csv", sep = ""))
}