Plot binary variable with value 1 across time-CodePudding

I have a dataset that includes a binary variable (0 vs 1) across time. I want to plot the occurrence of 1 across time. The idea is to see during which period the 1 occurs more often. Here is an example dataset:

set.seed(123)
    
dd <- data.frame(id = c(rep(1,100), rep(2,80), rep(3,90)),
                     time = c(seq(from=1,to=100), seq(from=1,to=80), seq(from=1,to=90)),
                     outcome = c(sample(c(0, 1), 100, replace = TRUE), 
                                 sample(c(0, 1), 80, replace = TRUE),  
                                 sample(c(0, 1), 90, replace = TRUE)))

I am thinking maybe a

My heatmap will have my time variable instead of Day for the x axis and the heat scale will reflect the frequency of 1. For my example dataset, because 1 occurs most often in the second period of time it will be highlighted as the most heated place in the plot.

CodePudding user response：

Not sure if I fully understand what you're looking for. Is it something like this:

library(tidyverse)
dd %>%
  group_by(time) %>% 
  summarise(outcome=sum(outcome)) %>% 
  ggplot(aes(x=time, fill=outcome))   geom_tile(aes(y=1))  
  scale_x_continuous(breaks = 1:7)   scale_fill_viridis_c(option='plasma')

It does highlight the second period as the highest frequency of 1 as required. If you need to split it into months (as per your example image) then use facet_wrap() or facet_grid() with the relevant month variable.

CodePudding user response：

Does something like this work?

dd %>% group_by(time) %>% summarise(n_outcome = sum(outcome))   %>%
  ggplot(aes(x = time, y = n_outcome))    geom_line()

You could obviously tweak it to make it prettier