Lexis lifeline ggplot

gglexis.R

library(ggplot2)
library(dplyr)
library(scales)

# SAMPLE DATA
set.seed(123)
# get all dates in a range
start_date <- as.Date("2017-01-01")
end_date <- as.Date("2017-06-01")
dates <- seq(start_date, end_date, 1)
# sample random start date with random duration 1-100
d <- data.frame(start = sample(dates, 100, replace=T),
                duration = round(runif(100, 1, 100)))
# end date as start date + duration
d$end <- d$start + d$duration

d <- d %>% 
  # infer whether active
  mutate(active = if_else(end >= end_date, 1, 0)) %>%
  # censor end date
  mutate(end = if_else(end > "2017-06-01", as.Date("2017-06-01"), end)) %>%
  # recalculate duration
  mutate(duration = end - start)

# GGPLOT
mytheme <- theme_bw() +
  theme(aspect.ratio=0.5,
        axis.title = element_text(size = 14),
        axis.text.y = element_text(size = 14),
        axis.text.x = element_text(angle = 45, hjust = 1, size = 13),
        legend.position = c(0, 1), 
        legend.justification = c(0, 1),
        legend.box.margin=margin(c(20,20,20,20)),
        legend.title = element_blank(),
        legend.text = element_text(size = 13),
        legend.background = element_rect(fill = "grey95"))

ggplot(d) +
  geom_segment(aes(x = start, y = 0, xend = end, yend = duration, colour = as.factor(active)), alpha = 0.3) +
  geom_point(aes(x = end, y = duration, colour = as.factor(active)), alpha = 0.6, size = 2) +
  scale_color_manual(name = "", labels = c("Inactive", "Active"), breaks = c(0, 1), values = c("grey30", "purple")) +
  scale_x_date(date_breaks = "1 month", date_minor_breaks = "1 month", labels = date_format("%b %y")) +
  scale_y_continuous(limits = c(0, 100), expand=c(0,0)) +
  xlab("") +
  ylab("Time (days)") +
mytheme