Exploratory data analysis

expl.R

@@ -1,9 +1,11 @@
 # Exploration on data
+# Export images 700x400
 
 # Libraries
 library(ggplot2)
 library(dplyr)
 library(RColorBrewer)
+library(scales)
 
 # Load data from Rdata file
 load("crashdata.Rdata")
@@ -38,7 +40,8 @@ vehtab
 
 # Graph crash severity
 ggplot(dat) + geom_bar(aes(x=CRASH_SEV), fill=NA, col="black") +
-  theme_classic() + labs(x = "Crash severity", y = "Number of crashes")
+  theme_classic() + labs(x = "Crash severity", y = "Number of crashes") +
+  scale_y_continuous(labels = scales::comma)
 
 # Not all openroad crashes are on state highways
 table(dat$URBAN, dat$CRASH_SH_DESC)
@@ -60,3 +63,9 @@ ggplot(rurhwy) +
                       labels = c("Single vehicle", "All other categories")) +
   scale_x_discrete("Severity", labels = c("Not", "Moderate", "Severe", "Fatal")) +
   theme_classic() + theme(legend.position = "right")
+
+rurtype <- as.data.frame.matrix(table(rurhwy$CRASH_SEV, rurhwy$MULTI_VEH))
+rurtypefrac <- rurtype / rowSums(rurtype)
+rurtypefrac["Single vehicle"]
+
+colSums(rurtype)/sum(rurtype)