Add days after a date in the x-axis

[carlanetto]

Hi! I need to create a timeline that shows the number of days after the start and not specific days in the x-axis. Is there a way of doing it? Thanks in advance Best regards

[shosaco]

I can't imagine what the result would look like, could you draw a picture or sketch?

[carlanetto]

It would look like any gantt chart, when you do not know yet when something will happen in time, just the order of events and number of days of each task:

https://images.app.goo.gl/SKx5D6EMuJedi4dP7

[shosaco]

The screenshot behind your link does show projects in the y-axis, not days, so kindly specify what you have tried and what is failing, thanks :)

[carlanetto]

The image says on the x-axis: “days of the month”. I wish to add numbers of days and not specific dates. Your package right now gives back an error and requested that the start and the end to be in a date format.

I want to add to the timeline: this project should start on day 6 and go until day 10 after the start of the project (that I don’t know when it will be). But without a month or year.

[carlanetto]

The image is of what I want to do and not an output of your package... Sorry, I thought that was clear

[shosaco]

So you are talking about the x-axis, not the y axis and already have tried something in code - please post this code as a starting point. The solution will be to convert your integer days into dates by using as.Date(2000-01-01) + your_days_vector.

[carlanetto]

I do not want it to show up like dates, so I don’t want to transform it do date. I just would like that the package does not require the input to be in a date format. This way the package could be applied in different contexts. But I will look for another solution in another package.

The code I am using is the example you give in the documentation.

Thanks for the attention

[carlanetto]

So you are talking about the x-axis, not the y axis and already have tried something in code - please post this code as a starting point. The solution will be to convert your integer days into dates by using as.Date(2000-01-01) + your_days_vector.

Thanks... indeed I am using it in x-axis, and I corrected it on the question. But it could actually be a nice idea to be able to flip the visualization.

[shosaco]

I used my suggestions and formatted the x-axis accordingly:

data <- read.csv(text="event,start,end
                       Phase 1,1,10
                       Phase 2,8,15", stringsAsFactors=FALSE)
data$start <- as.Date("2000-01-01") + data$start - 1
data$end <- as.Date("2000-01-01") + data$end - 1
gg_vistime(data) +
    scale_x_datetime(breaks = seq(min(as.POSIXct(data$start)), 
                                  max(as.POSIXct(data$end)), 
                                  "days"), 
                     date_labels = "%d") + 
    xlab("Day")

[jonocarroll]

This took me a while to figure out so here it is for anyone else searching...

The above solution works great for a static ggplot but doesn't work if you want to keep the plotly interactivity, which I did.

I re-worked the data example; in this case the start and end represent offsets from the start date. This is perhaps more common for 'days on study' or 'days since an significant event'

library(vistime)
## start and end represent days from 2000-01-01
## so starts are days 2 and 9, ends are days 11 and 16
data <- read.csv(text="event,start,end
                       Phase 1,1,10
                       Phase 2,8,15", stringsAsFactors=FALSE)
epoch <- as.POSIXct("2000-01-01", tz = "UTC")
data$start <- epoch + lubridate::days(data$start)
data$end <- epoch + lubridate::days(data$end)

Your example above still works (I shifted this to reflect 'days after start')

gg_vistime(data) +
  ggplot2::scale_x_datetime(breaks = seq(epoch, max(data$end), "days") - 1, 
                            date_labels = "%d") + 
  ggplot2::xlab("Day") 

But by default, vistime() uses the dates

vistime(data)

The x-axis tick-labels can be specified in plotly using layout()

daysrange <- seq(epoch, to = max(data$end), "days")
ndays <- length(daysrange)

vistime(data) %>%
  plotly::layout(xaxis = list(title = "Offset Day", 
                              tickmode = "array",
                              tickvals = daysrange,
                              ticktext = 0:ndays))

This also opens up the possibility of combining the two with custom labels

daystext <- paste0(0:ndays, "<br>(", format(daysrange, "%b %d"), ")")
daystext[1] <- sub(")", paste0(", ", lubridate::year(epoch), ")"), daystext[1])
vistime(data) %>%
  plotly::layout(xaxis = list(title = "Offset Day", 
                              tickmode = "array",
                              tickvals = daysrange[c(TRUE, FALSE)],
                              ticktext = daystext[c(TRUE, FALSE)]))

Thanks for this package, it's very useful!