inria phishing plot.R

library(tidyverse)

data = tribble(
  ~Site,	~Participants,	~Victims,	~Percent_Victims,
  "Bordeaux",	137,	12,	8.8,
  "Lille",	131,	9,	6.9,
  "Lorraine",	268,	37,	13.8,
  "Paris",	239,	23,	9.6,
  "Rennes",	324,	33,	10.2,
  "Rhone-Alpes",	219,	13,	5.9,
  "Rocquencourt (Siege)",	82,	8,	9.8,
  "Saclay",	204,	7,	3.4,
  "Sophia",	219,	28,	12.8
)

data %>% 
  rowwise() %>% 
  mutate( lo = PropCIs::exactci(Victims, Participants, .95)$conf.int[[1]],
          hi = PropCIs::exactci(Victims, Participants, .95)$conf.int[[2]],
          lo99 = PropCIs::exactci(Victims, Participants, .99)$conf.int[[1]],
          hi99 = PropCIs::exactci(Victims, Participants, .99)$conf.int[[2]],
          lo50 = PropCIs::exactci(Victims, Participants, .5)$conf.int[[1]],
          hi50 = PropCIs::exactci(Victims, Participants, .5)$conf.int[[2]],
          Percent_Victims = Victims / Participants) %>% 
  ungroup() %>%
  arrange(lo) %>% 
  mutate(Site = fct_inorder(Site, ordered = TRUE)) %>% 
  ggplot() +
  aes(x = Site, y = Percent_Victims) +
  geom_point(size = 4) +
  geom_segment(aes(x=Site, xend=Site, y=lo, yend=hi), size = 1) +
  geom_segment(aes(x=Site, xend=Site, y=lo50, yend=hi50), size = 2) +
  geom_segment(aes(x=Site, xend=Site, y=lo99, yend=hi99), size = 0.33) +
  scale_y_continuous(labels = function(y) paste0(y*100, "%")) +
  labs(x = NULL, y = NULL, title = "Inria phishing victims") +
  coord_flip() +
  theme_light(20)