PCA and k-means data

library(tidyverse)
library(factoextra)

agg_transp <- read_csv("data/aggregated_transport.csv")

transp_border <- agg_transp %>% 
  group_by(location) %>% 
  mutate(mean_value = mean_value/max(mean_value))


pca_cases <- transp_border %>% 
  select(date, mean_value, location) %>% 
  ungroup() %>% 
  spread(key = date, value = mean_value, fill = 0) %>% 
  select(-location) %>% 
  as.matrix()

# PCA with this matrix
pca_cases <- pca_cases[ , which(colSums(pca_cases) > 0 )]
pca_m0 <- prcomp(pca_cases, center = TRUE, scale = TRUE)

resp <- transp_border %>% 
  group_by(location) %>% slice(1) %>% 
  pull(border_station)


fviz_pca_ind(pca_m0, geom.ind = "point", pointshape = 21, 
             pointsize = 2, 
             fill.ind = factor(resp), 
             col.ind = "black", 
             palette = "jco", 
             addEllipses = TRUE,
             label = "var",
             col.var = "black",
             repel = TRUE,
             legend.title = "Bordering county") +
  ggtitle("2D PCA-plot") +
  theme_classic(18) +
  theme(plot.title = element_text(hjust = 0.5), 
        legend.position = "bottom")

# k-means
km <- kmeans(pca_cases, centers = 3)
clusters <- km$cluster

fviz_pca_ind(pca_m0, geom.ind = "point", pointshape = 21, 
             pointsize = 2, 
             fill.ind = factor(clusters), 
             col.ind = "black", 
             palette = "jco", 
             addEllipses = TRUE,
             label = "var",
             col.var = "black",
             repel = TRUE,
             legend.title = "Bordering county") +
  ggtitle("2D PCA-plot") +
  theme_classic(18) +
  theme(plot.title = element_text(hjust = 0.5), 
        legend.position = "bottom")
#-----------------------------------------------------------------------