# Load chemRmetrics package - If not installed, refer to the guide file in the ChemRmetrics repository on GitHub
library(chemRmetrics)

# Load data from jdx files in 'Market Survey Raw Data' folder
raw_data <- load_data('Markey Survey Raw Data', ftype = 'jdx', fstruc = 'Brand_Filament_Colour_Form_Replicate', delim = '_')

# Remove diamond signal from all spectra
trunc <- data_trunc(raw_data, 6, 2340, 1850)

# Interpolate the data between a range of 4000-400 cm-1, with a step size of 1
interp <- data_interp(trunc, 6, 4000, 400, 1)

# Normalise the spectra to unity
norm <- data_norm_mm(interp, 6)

# Perform principal component analysis (also exports scree plot and loadings plots for first 7 PCs)
pc_data <- pca(norm, 6, 7)

# Create a 3D PCA scores plot using PC1, PC2, and PC3 and the 'Filament' variable for the colour and legend (marker size = 3)
plot_3D(pc_data, 'Filament', 1, 2, 3, 3, c('#A75529', '#5DD291', '#8B0069'))

# The following for-loop was used to adjust filament-type to be just PLA, ABS, and PETG (e.g. PLA+ -> PLA, PETG-CF -> PETG ...etc)
for (i in 1:nrow(raw_data)){
  if (raw_data[i, 3] == 'PLA+'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Flex'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Wood'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Rainbow'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-CF'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Tough'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Bicolour'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Copper'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Marble'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'PLA-Twinkling'){
    raw_data[i, 3] <- 'PLA'
  } else if (raw_data[i, 3] == 'ABS+'){
    raw_data[i, 3] <- 'ABS'
  } else if (raw_data[i, 3] == 'PETG-CF'){
    raw_data[i, 3] <- 'PETG'
  }
}

# The remainder of the script is repeat of the above data analysis using the simplified PLA, ABS, and PETG labels
# to investigate the grouping of the main polymer classes

trunc_2 <- data_trunc(raw_data, 6, 2340, 1850)

interp_2 <- data_interp(trunc_2, 6, 4000, 400, 1)

norm_2 <- data_norm_mm(interp_2, 6)

pc_data_2 <- pca(norm_2, 6, 7)

pc_data_cut <- pca(norm_cut, 6, 7)

plot_3D(pc_data_cut, 'Filament', 1, 2, 3, 3, colours = c('#A75529', '#5DD291', '#8B0069'))