Illustrating relative timeline in R

I have a dataframe with two columns: los that denotes length of stay for a particular event and daysbetweenevent. I would like to create a visualization that shows a single line with segments illustrating width of los and daysbetween event alternating. I'm thinking a thick band for los and a thin band for daysbetweenvisits. And each segment would take up the relative length for the number - essentially showing a timeline.

For example, the order of the segments would be: 2-21-24-12-14-52...., with every los segment being thick, and every daysbetweenvisits being a line, with preferable label of the number on top.

df <- data.frame(los = c(2, 24, 14, 0, 4, 9, 7, 3, 1), 
                 daysbetweenvisits = c(21, 12, 52, 218, 73, 36, 0, 18, 0))

The first step is getting your data into the correct format. For this, you want to calculate the start of each time period and end of each time period, then label it according to whether it is during the 'LOS' period or the 'between visits' period.

You can do this by transposing your data frame and concatenating it into a single numeric vector. Then its cumulative sum will be the time points at the end of each segment. Then lag this vector with a 0 in the first position to get the start time. You can put these in a data frame with a second column that just alternates between 'LOS' and 'BetweenVisits':

df <- data.frame(right = cumsum(c(t(df))), 
                 left = head(c(0, cumsum(c(t(df)))), -1),
                 event = rep(c('Stay', 'Between'), len = nrow(df) * 2))
df$right <- ifelse(df$right == df$left, df$right + 1, df$right)

So now the data is in the correct format for plotting:

df
#>    right left   event
#> 1      2    0    Stay
#> 2     23    2 Between
#> 3     47   23    Stay
#> 4     59   47 Between
#> 5     73   59    Stay
#> 6    125   73 Between
#> 7    126  125    Stay
#> 8    343  125 Between
#> 9    347  343    Stay
#> 10   420  347 Between
#> 11   429  420    Stay
#> 12   465  429 Between
#> 13   472  465    Stay
#> 14   473  472 Between
#> 15   475  472    Stay
#> 16   493  475 Between
#> 17   494  493    Stay
#> 18   495  494 Between

And we can draw the plot with ggplot2

library(ggplot2) 

ggplot(df, aes(left, '1', linewidth = event, group = event, color = event)) +
  geom_segment(aes(xend = right, yend = '1')) +
  geom_text(aes(left + (right - left)/2, 
                label = ifelse(event == 'Between', right - left, '')), 
            nudge_y = 0.1, color = 'black') +
  scale_linewidth_manual(values = c(1, 6), guide = 'none') +
  scale_color_manual(values = c('gray', 'red3'), guide = 'none') +
  theme_void()

^{Created on 2023-07-27 with reprex v2.0.2}

Here's base R approach:

# find LOS endpoints
df$from <- c(0, cumsum(head(rowSums(df), -1)))
df$to <- df$from + df$los

par(mar=rep(0, 4))
frame()
plot.window(xlim=c(0, tail(df$to, 1)), ylim=c(0, 0))
segments(df$from[1], 0, tail(df$to, 1), lend=3)
segments(df$from, 0, df$to, lwd=16, col='red', lend=3)
text((head(df$to, -1) + df$from[-1]) / 2, 2*strheight('a'), 
     head(df$daysbetweenvisits, -1))

I wasn't sure how to interpret the zero values so I just kept them zero.