plotting running time distribution from

lundaloppet_plot.R

library(readxl)
library(tidyverse)
library(hms)
library(hrbrthemes)
library(extrafont)

read_excel("lundaloppet2018.xlsx", col_names = F, col_types = "text") %>% 
  janitor::clean_names() %>% 
  mutate(x2=parse_number(x2),
         x5n = parse_number(x5),
         x5c = factor(str_remove_all(x5, "\\d+"), 
                      levels = c("T", "P"), 
                      #labels=c("Women", "Men")
                      labels=c("Damer", "Herrar")
                      ),
         x14=parse_number(x14),
         x15=parse_number(x15),
         x16=parse_number(x16)) %>% 
  drop_na(x2) %>% 
  filter(x5n==5) %>% 
  mutate(time=x14*60*60+x15*60+x16) %>% {
  ggplot(., aes(x=time, fill=x5c, color=x5c))+
  #geom_histogram(bins=60)+
  geom_density(alpha=0.3)+
  #stat_ecdf(size=1.2)+
  scale_x_time(breaks = seq(10*60, 120*60, by=10*60), expand=c(0.1, 0))+
  theme_ipsum_rc(grid=FALSE)+
  scale_fill_ipsum()+
  scale_color_ipsum()+
  #facet_wrap(vars(x5))+
  labs(
       title= "Resultat för Lundaloppet 2018",
       subtitle= paste("Fördelning för löptider för", first(.$x5n),"km"),
       caption="Källa: www.lundaloppet.se",
       x="Löptider (hh:mm:ss)",
       
       #title="Results for Lundaloppet 2018",
       #subtitle= paste("Distribution of running times for", first(.$x5n),"km"),
       #caption="Source: www.lundaloppet.se",
       #x="Running time (hh:mm:ss)",

       y=NULL, fill=NULL, color=NULL)+
  theme(axis.text.y = element_blank(),
        legend.position = c(0.9, 0.3))
}