rm(list=ls())
library(dplyr)
require(ggplot2)
require(plotly)
require(heatmaply)
require(factoextra)

dados =  dados[dados$MODALIDADE == "PMPL_FALL_2023",]

sum(is.na(dados))

data = mutate_all(dados, ~replace(., is.na(.), 0))
sum(is.na(data))

sel = c()
for(i in 1:ncol(data)){
  if(class(data[,i]) == 'numeric' | class(data[,i]) == 'integer'){
    sel = append(sel,i)
  }
}
data = data[,sel]
colnames(data)

X = data
round(sapply(X, var),1)

X = X[,sapply(X, var) != 0]
sel = c()
for(i in 1:ncol(X)){
  if(dim(table(X[,i] < 0)) == 2){
    sel = append(sel,i)
  }
}
if(length(sel) > 0){
  X = X[,-sel]
} else{
  
}

excluir = c("Queda","Dia","Dano_1")
X <- X[, setdiff(names(X), excluir)]

{

report = c("Abates"     ,            "Assistencias"    ,       "Cura"  ,                 "Dano"           ,        "Granada.Fumaça"  ,      
  "Headshots"       ,       "Abates.por.granada" ,    "Knockouts"   ,           "Tempo.sobrevivencia"   , "INTENSIDADE_DE_COMBATE",
 "DMG_P_CURA" ,            "PART_EFET")

report2 = c("Abates"     ,            "Assistencias"    ,       "Cura"  ,                 "Dano"           ,        "Granada.Fumaça"  ,      
  "Headshots"       ,       "Abates.por.granada" ,    "Knockouts"   ,           "Tempo.sobrevivencia"   , "INTENSIDADE_DE_COMBATE",
 "DMG_P_CURA" )
}

#X1 = scale(X[,c(coletivo,individual,recursos,estrategia,vs,objetivos)],center = T,scale = T)
X1 = X[,c(report2)]
acpcor <- princomp(X1, cor= T)
screeplot(acpcor, type = c("lines"))

mat = (acpcor$loadings[,1:5])

p <- heatmaply(round(mat,2), 
        dendrogram = "none",
        xlab = "", ylab = "", 
        main = "",
        scale = "column",
        margins = c(60,100,40,20),
        grid_color = "white",
        grid_width = 0.00001,
        titleX = FALSE,
        hide_colorbar = TRUE,
        branches_lwd = 0.1,
        #labCol = colnames(mat),
        #labRow = rownames(mat),
        heatmap_layers = theme(axis.line=element_blank()),
        col = colorRampPalette(c("#FFCC00","#FF9900",
                                 "#FF6600","#FF3300","#FF0000"))(100),
        na.rm = T,cellnote = round(mat,2)
        ,cellnote_size = 10,cellnote_textposition = "middle center")
p

dados$comp1 = acpcor$scores[,1]
dados$comp2 = acpcor$scores[,2]
dados$comp3 = acpcor$scores[,3]

dados[sample(1:nrow(dados),15),c('Jogador','comp1','comp2','comp3')]

avg_comp = dados |> group_by(Jogador) |> select(comp1,comp2,comp3) |> summarise_all(mean)
qtd_part = dados |> group_by(Jogador) |> summarise(qtd = n())

players = merge(avg_comp,qtd_part,by = c('Jogador'))

sel = players |> filter(qtd >= quantile(players$qtd,.5)) |> top_n(10,comp1) %>% select(Jogador,comp1) |> arrange(-comp1)

ggplot(data = sel, mapping = aes(x = reorder(Jogador,+comp1), y = comp1, fill = Jogador))+
  labs(title = "comp1")+
  xlab("")+ ylab("")+
  geom_bar(stat = "identity")+
  theme(legend.position = "none",panel.background = element_rect(fill = "#17202A"),plot.title = element_text(hjust = 0.5),axis.text.x = element_blank(), panel.grid = element_blank())+
  geom_text(aes(label = round(comp1,3)), vjust = 0.5,hjust = 1,color = "white")+
  coord_flip()

sel = players |> filter(qtd >= quantile(players$qtd,.5)) |> top_n(10,comp2) %>% select(Jogador,comp2) |> arrange(-comp2)

ggplot(data = sel, mapping = aes(x = reorder(Jogador,+comp2), y = comp2, fill = Jogador))+
  labs(title = "comp2")+
  xlab("")+ ylab("")+
  geom_bar(stat = "identity")+
  theme(legend.position = "none",panel.background = element_rect(fill = "#17202A"),plot.title = element_text(hjust = 0.5),axis.text.x = element_blank(), panel.grid = element_blank())+
  geom_text(aes(label = round(comp2,3)), vjust = 0.5,hjust = 1,color = "white")+
  coord_flip()

sel = players |> filter(qtd >= quantile(players$qtd,.5)) |> top_n(10,comp3) %>% select(Jogador,comp3) |> arrange(-comp3)

ggplot(data = sel, mapping = aes(x = reorder(Jogador,+comp3), y = comp3, fill = Jogador))+
  labs(title = "comp3")+
  xlab("")+ ylab("")+
  geom_bar(stat = "identity")+
  theme(legend.position = "none",panel.background = element_rect(fill = "#17202A"),plot.title = element_text(hjust = 0.5),axis.text.x = element_blank(), panel.grid = element_blank())+
  geom_text(aes(label = round(comp3,3)), vjust = 0.5,hjust = 1,color = "white")+
  coord_flip()

dt = dados
row.names(dt) = dt[,1]
dt = dt[,-1]
df <- scale(dt)
df.dist = dist(df, method = "euclidean")

df.hclust = hclust(df.dist, method = "ward.D2")
plot(df.hclust)

p = fviz_dend(df.hclust, k = 15,
          
          cex = .7, 
          #k_colors = c("#2E9FDF", "#c066c0", "#E7B800", "#FC4E07"),
          color_labels_by_k = TRUE, 
          rect = TRUE,
          #rect_border = c("#2E9FDF", "#cc84cc", "#E7B800", "#FC4E07"),
          rect_fill = TRUE,
          #horiz = T,
          main = "",
          type = "rectangle"
          ,ylab = "",horiz = T
          )
p

groups.2 = cutree(df.hclust,15)
dados <- cbind(dados, clusterg2 = groups.2)

mat = dados %>% select(-Jogador) %>%
  group_by(clusterg2) %>%
  summarize_all(mean) %>% select(-clusterg2)


mat = mat %>% mutate_all(~ replace(., is.nan(.), NA))


mat = mat[,which(!(is.na(sapply(mat, sd)) | sapply(mat, sd) == 0))]

#sort(sapply(mat, var),decreasing = F)

mat = round(mat[,-which(colnames(mat) %in% c('Headshots','Abates.por.granada'))],2)


p <- heatmaply(mat[,1:round(ncol(mat)/2,0)], 
        dendrogram = "none",
        xlab = "", ylab = "Clusters", 
        main = "",
        scale = "column",
        margins = c(60,100,40,20),
        grid_color = "white",
        grid_width = 0.00001,
        titleX = FALSE,
        hide_colorbar = TRUE,
        branches_lwd = 0.1,
        #labCol = colnames(mat),
        #labRow = rownames(mat),
        heatmap_layers = theme(axis.line=element_blank()),
        col = colorRampPalette(c("#FFCC00","#FF9900",
                                 "#FF6600","#FF3300","#FF0000"))(100),
        na.rm = T,cellnote = round(mat,2)
        ,cellnote_size = 10,cellnote_textposition = "middle center")
p

p <- heatmaply(mat[,round(ncol(mat)/2,0):ncol(mat)], 
        dendrogram = "none",
        xlab = "", ylab = "Clusters", 
        main = "",
        scale = "column",
        margins = c(60,100,40,20),
        grid_color = "white",
        grid_width = 0.00001,
        titleX = FALSE,
        hide_colorbar = TRUE,
        branches_lwd = 0.1,
        #labCol = colnames(mat),
        #labRow = rownames(mat),
        heatmap_layers = theme(axis.line=element_blank()),
        col = colorRampPalette(c("#FFCC00","#FF9900",
                                 "#FF6600","#FF3300","#FF0000"))(100),
        na.rm = T,cellnote = round(mat,2)
        ,cellnote_size = 10,cellnote_textposition = "middle center")
p

dados[dados$clusterg2 == 1,]$Jogador

dados[dados$clusterg2 == 4,]$Jogador

dados[dados$clusterg2 == 5,]$Jogador

dados[dados$clusterg2 == 15,]$Jogador