boxplot

projekt.R

@@ -43,3 +43,28 @@ ggplot(german_credit_risk , aes(x = factor(Job), fill = Purpose)) +
   scale_x_discrete(breaks = 0:3, labels=c("Unskilled, non-resident", "Unskilled, resident","Skilled","Highly skilled")) +
   labs(title="Credit count and purpose for different job statuses", x="Job status", y = "Credit count") +
   theme_minimal()
+
+#Wykres pudełkowy przedstawiający jakiej wielkości  najczęściej ludzie biorą kredyty na poszczególne cele.
+stat_box_data <- function(y) {
+  return( 
+    data.frame(
+      y = max(german_credit_risk$Credit.amount),
+      label = paste('count =', length(y), '\n',
+                    'sum =', sum(y), '\n')
+    )
+  )
+}
+
+german_credit_risk %>%
+  mutate(Purpose = reorder(Purpose, Credit.amount, sum)) %>%
+  ggplot(aes(x = Purpose, y = Credit.amount, fill = Purpose)) +
+  geom_boxplot(alpha = .7) +
+  guides(fill = "none") +
+  scale_y_continuous(breaks = scales::pretty_breaks(n = 10)) +
+  stat_summary(
+    fun.data = stat_box_data,
+    geom = "text",
+    vjust = 1
+  ) +
+  ylab("Credit amount") +
+  theme_bw()