# Skript slouzi k:
# 1. Nacteni souboru s daty (Data_neuro.xlsx).
# 2. Uprave hodnot kategorialnich promennych.
# 3. Vizualizaci dat:
#    a) Teckovy graf podle kategorialni promenne
#    b) Maticovy graf
#    c) Maticovy graf podle kategorialni promenne
#    d) Krabicove grafy pro vice spojitych promennych
#    e) Vicenasobne krabicove grafy (pro vice kategorialnich promennych a jednu spojitou promennou)
#    f) Ikonove grafy 
# Vytvoreno 27.2.2019; EK


##### Nacteni knihoven pro praci s daty #####
library(readxl)   # nacteni knihovny s funkci read_excel
library(plyr)     # nacteni knihovny s funkci mapvalues
library(aplpack)  # nacteni knihovny s funkci faces


##### Nacteni datoveho souboru #####

wd <- "c:/Users/koritakova/Documents/Vyuka/DSAN02/2019 jaro/DSAN02_predn01_Uvod/" # cesta ke slozce, ve ktere je datovy soubor

data1 <- read_excel(file.path(wd,"Data_neuro.xlsx"),sheet="data") # nacteni dat

dim(data1) # pocet pacientu a pocet promennych
names(data1) # nazvy promennych


##### Uprava dat #####

# Vytvoreni promenne s textovymi popisky skupin subjektu
data1$Group_3kat_rek <- mapvalues(data1$Group_3kat,from=levels(as.factor(data1$Group_3kat)),to=c("CN","MCI","AD")) # vytvoreni nove promenne s textovymi hodnotami kategorii
data1$Group_3kat_rek <- factor(data1$Group_3kat_rek, levels=c("AD","MCI","CN")) # prerazeni poradi kategorii
table(data1$Group_3kat_rek) # pocet subjektu v kazde kategorii

# Vytvoreni kategorizovaneho veku:
data1$Age_kat <- cut(data1$Age, c(0,seq(70,80,10),200), right=FALSE)
levels(data1$Age_kat) <- c("<70","70-79","80+") # rename labels
table(data1$Age_kat)

# Prevedeni promenne Gender_rek na faktor
data1$Gender_rek <- as.factor(data1$Gender_rek)
table(data1$Gender_rek)


##### Vizualizace dat #####

## a) Teckovy graf podle kategorialni promenne

barvy = c('green','orange','red') # nadefinovani barev pro kategorialni promennou
plot(data1$Hippocampus_volume_mm3, data1$Amygdala_volume_mm3, pch=20, col=barvy[as.numeric(data1$Group_3kat)], xlab="Hippocampus volume (mm3)", ylab="Amygdala volume (mm3)") # vykresleni grafu (pch - typ symbolu, col - barva, xlab - popisek na ose x, ylab - popisek na ose y)
legend("bottomright", c("CN","MCI","AD"), pch=20, col=barvy, cex=0.9, bty="n") # pridani legendy (pch - typ symbolu, col - barva, cex - velikost popisku, bty - zobrazeni obdelniku kolem legendy ("n" = ne))


## b) Maticovy graf

pairs(~Age+Height+Hippocampus_volume_mm3+Amygdala_volume_mm3, col="black", data=data1) # vykresleni grafu

pairs(~Age+Height+Hippocampus_volume_mm3+Amygdala_volume_mm3, panel= function(x,y,...) {
  points(x, y,...)
  abline(lm(y~x),col="grey")}, col="black", data=data1) # vykresleni grafu vcetne pridani primky linearni regrese


## c) Maticovy graf podle kategorialni promenne

barvy = c('green','orange','red') # nadefinovani barev pro kategorialni promennou
pairs(~Age+Height+Hippocampus_volume_mm3+Amygdala_volume_mm3,col=barvy[as.numeric(data1$Group_3kat)], data=data1) # vykresleni grafu


## d) Krabicove grafy pro vice spojitych promennych

par(mar=c(12.1, 4.1, 4.1, 2.1)) # zvetseni prostoru pod grafem, aby se tam vesly popisky
boxplot(data1[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3","Nucl_caud_volume_mm3")],xaxt='n',las=2)  # xaxt='n' - nevykresli popisky na ose x, las=2 - otoci popisky na ose y
axis(1, 1:6, c("Hippocampus volume (mm3)","Amygdala volume (mm3)","Thalamus volume (mm3)","Pallidum volume (mm3)","Putamen volume (mm3)","Nucl caud volume (mm3)"), las=2) # vykresleni otocenych popisku na ose x 

dev.off() # zavreni grafickeho okna


## e) Vicenasobne krabicove grafy (pro vice kategorialnich promennych a jednu spojitou promennou)

table(data1$Group_3kat_rek,data1$Gender_rek) # pomocna kontingencni tabulka pro predstavu, kolik bude subjektu v jednotlivych grafech
n_radku <- length(levels(data1$Group_3kat_rek)) # pocet radku s grafy
n_sloupcu <- length(levels(data1$Gender_rek)) # pocet radku s grafy
n_boxu <- length(levels(data1$Age_kat)) # pocet krabicovych grafu v 1 grafu

par(mfrow=c(n_radku,n_sloupcu)) # nastaveni "gridu" grafu
par(oma=c(3,3,3,3))     # nastaveni "outer margin area"
par(mar=c(1,1,1,1))     # nastaveni "margin" u jednotlivych grafu 
par(mgp = c(0, 0.5, 0)) # nastaveni popisku blize "tickmarks"
for (j in 1:n_radku) {
  for (i in 1:n_sloupcu) {
    data1_sel <- data1[data1$Group_3kat_rek==levels(data1$Group_3kat_rek)[j] & data1$Gender_rek==levels(data1$Gender_rek)[i],] # vyber dat podle j-te kategorie 1. promenne a i-te kategorie 2. promenne
    if (nrow(data1_sel)>0) {
      bstat=boxplot(data1_sel$Hippocampus_volume_mm3 ~ data1_sel$Age_kat, range=0, plot=FALSE) # vypocet charakteristik (minima, 1.kvartilu, medianu, 2.kvartilu, maxima)
      plot(1:n_boxu, bstat$stats[3,], col="blue", pch=18, cex=2, xaxt="n", xlim=c(0.5,n_boxu+0.5), ann=FALSE, ylim=c(5500,7500),cex.axis=0.9) # vykresleni medianu (xaxt="n" zabrani vykresleni osy x) 
      arrows(1:n_boxu, bstat$stats[1,], 1:n_boxu, bstat$stats[5,], code = 3, col = "blue", angle = 90, length = .1) # vykresleni minima a maxima
      axis(1,at=c(1:n_boxu),labels=levels(data1$Age_kat),cex.axis=0.9) # vykresleni osy x
      for (k in 1:n_boxu) {
        text(k,11,paste("n=",length(data1_sel$Hippocampus_volume_mm3[data1_sel$Age_kat==levels(data1_sel$Age_kat)[k]]),sep=""),cex=0.9) # pridani textovych popisku "n=XX"
      }
    } else {
      plot(1,1,col="white",xaxt="n",yaxt="n",axes=F,frame.plot=FALSE,ann=FALSE) # v pripade, ze pro danou kombinaci kategorii nejsou data zadneho pacienta, vykresli se prazdny graf
    }
  }
}
for (j in 1:n_radku) {
  mtext(levels(data1$Group_3kat_rek)[j],side=2,line=1,outer=TRUE,at=1.17-j/n_radku,cex=0.8) # popisky pro kategorialni promennou v radcich
}
for (i in 1:n_sloupcu) {
  mtext(levels(data1$Gender_rek)[i],side=3,line=1,outer=TRUE,at=i/n_sloupcu-0.25,cex=0.75) # popisky pro kategorialni promennou ve sloupeccich
}

dev.off() # zavreni grafickeho okna


## f) Ikonove grafy 

data1_sel=data1[c(1:10,801:810),] # vyber 20 subjektu

stars(data1_sel[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3")], labels=data1_sel$ID, ncol=5) # parskove (hvezdicove) grafy
stars(data1_sel[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3")], labels=data1_sel$ID, ncol=5, radius=F, col.stars=rep("blue",length(data1_sel$ID))) # polygony

faces(data1_sel[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3")], face.type=0) # Chernoffovy tvare - typ 0
faces(data1_sel[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3")], face.type=1) # Chernoffovy tvare - typ 1
faces(data1_sel[,c("Hippocampus_volume_mm3","Amygdala_volume_mm3","Thalamus_volume_mm3","Pallidum_volume_mm3","Putamen_volume_mm3")], face.type=2) # Chernoffovy tvare - typ 2