Forcing the Time Axes to be set at a specific boundaries

Question

Here is what I have now:

graph_POST %>%
  mutate(day = lubridate::day(Date)) %>%
  pivot_longer(Standing:New_Sitting, names_to = "Posture") %>%
ggplot(aes(x = Time, y = value, color = Posture))+
  geom_point(size=0.5)+
  scale_y_continuous(limits = c(0.5,4.5), expand = expansion(0)) +
  facet_wrap(~day, scales = "free_x") +
  labs(title = "Posture vs. Time (POST)") +
  theme(axis.title.y = element_blank(),
        axis.text.y = element_blank(),
        panel.grid.major.y = element_blank(),
        panel.grid.minor.y = element_blank())

I need the x axes to be the same for all my plots in my "facet" (8am-5pm) with a tick every hour. I've tried to do similar to my y axes, but it seems to not like that for some reason:scale_x_continuous(limits = c(8:00,5:00))

My Output so far:

Please let me know if you have any questions or need clarifications

> dput(head(graph_POST,30))
structure(list(Date = structure(c(19145, 19145, 19145, 19145, 
19145, 19145, 19145, 19145, 19145, 19145, 19145, 19145, 19145, 
19145, 19145, 19145, 19145, 19145, 19145, 19145, 19145, 19145, 
19145, 19145, 19145, 19145, 19145, 19145, 19145, 19145), class = "Date"), 
    Time = structure(c(1654182900, 1654182901, 1654182902, 1654182903, 
    1654182904, 1654182905, 1654182906, 1654182907, 1654182908, 
    1654182909, 1654182910, 1654182911, 1654182912, 1654182913, 
    1654182914, 1654182915, 1654182916, 1654182917, 1654182918, 
    1654182919, 1654182920, 1654182921, 1654182922, 1654182923, 
    1654182924, 1654182925, 1654182926, 1654182927, 1654182928, 
    1654182929), class = c("POSIXct", "POSIXt"), tzone = ""), 
    Axis1 = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), Axis2 = c(0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0), Axis3 = c(0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0), VM = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), Standing = c(0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0), Stepping = c(0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0), Cycling = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
    0, 0), New_Sitting = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1), Counter = c(0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L)), row.names = c(NA, 
30L), class = "data.frame")

EDIT

graph_POST %>%
  mutate(day = lubridate::day(Date)) %>%
  pivot_longer(Standing:New_Sitting, names_to = "Posture") %>%
  ggplot(aes(x = Time, y = value, color = Posture)) +
  geom_point(data = data.frame(
    Time = as.POSIXct(rep(unique(graph_POST$Date), 2)) +
           rep(c(8, 17) * 3600, each = length(unique(graph_POST$Date))),
    value = 1, Posture = "Cycling", 
    day = rep(lubridate::day(unique(graph_POST$Date)), 2)), alpha = 0) +
  geom_point(size=0.5)+
  scale_y_continuous(limits = c(0.5,4.5), expand = expansion(0)) +
  scale_x_datetime(date_labels = "%H:%M",
                   breaks = rep(as.POSIXct(unique(graph_POST$Date)), 4) +
                     rep(c(8, 11, 14, 17) * 3600, 
                         each = length(unique(graph_POST$Date)))) +
  facet_wrap(~day, scales = "free_x") +
  labs(title = "Posture vs. Time (POST)") +
  theme(axis.title.y = element_blank(),
        axis.text.y = element_blank(),
        panel.grid.major.y = element_blank(),
        panel.grid.minor.y = element_blank())

for hourly scales, follow for example: https://stackoverflow.com/q/50568227/7941188 (needs only midl adjustments to *hourly* breaks — tjebo, Jun 13 '22 at 12:43
Removing Free_x just makes my plots very small since my data spans over multiple days, I Need them to be scaled from 8am-5pm for the specific date of that plot within the facet — Richard Michaud Langis, Jun 13 '22 at 12:59
Adding the scales argument of breaks clutters the x axes and makes it unreadable — Richard Michaud Langis, Jun 13 '22 at 13:00

score 3 · Accepted Answer · answered Jun 13 '22 at 13:57

3

You can use the old "add invisible points at 8am and 5pm in each facet" trick. This is a bit of a pain in this case because the faceting variable is completely correlated with the x axis. The code is also a bit longer than it had to be because I thought it would be better to put labels at exactly 08:00 and 17:00.

Note that your sample data only included a single day's data which means the problem isn't reproducible. I have therefore created a compatible data set for illustration (see below)

graph_POST %>%
  mutate(day = lubridate::day(Date)) %>%
  pivot_longer(Standing:New_Sitting, names_to = "Posture") %>%
  ggplot(aes(x = Time, y = value, color = Posture)) +
  geom_point(data = data.frame(
    Time = as.POSIXct(rep(unique(graph_POST$Date), 2)) +
           rep(c(8, 17) * 3600, each = length(unique(graph_POST$Date))),
    value = 1, Posture = "Cycling", 
    day = rep(lubridate::day(unique(graph_POST$Date)), 2)), alpha = 0) +
  geom_point(size=0.5)+
  scale_y_continuous(limits = c(0.5,4.5), expand = expansion(0)) +
  scale_x_datetime(date_labels = "%H:%M",
                   breaks = rep(as.POSIXct(unique(graph_POST$Date)), 4) +
                     rep(c(8, 11, 14, 17) * 3600, 
                         each = length(unique(graph_POST$Date)))) +
  facet_wrap(~day, scales = "free_x") +
  labs(title = "Posture vs. Time (POST)") +
  theme(axis.title.y = element_blank(),
        axis.text.y = element_blank(),
        panel.grid.major.y = element_blank(),
        panel.grid.minor.y = element_blank())

Reproducible data

set.seed(1)

graph_POST <- data.frame(Date = rep(as.Date(
  c("2022-06-02", "2022-06-03", "2022-06-06", "2022-06-07", "2022-06-08")),
  each = 100),
  Time = as.POSIXct(rep(as.Date(
  c("2022-06-02", "2022-06-03", "2022-06-06", "2022-06-07", "2022-06-08")),
  each = 100)) + sample(32000:61200, 500),
  Standing = sample(c(0, 4), 500, TRUE, prob = c(4, 1)),
  Stepping = sample(c(0, 3), 500, TRUE, prob = c(4, 1)),
  Cycling = sample(c(0, 2), 500, TRUE, prob = c(4, 1)),
  New_Sitting = sample(0:1, 500, TRUE, prob = c(4, 1)))

answered Jun 13 '22 at 13:57

Allan Cameron

147,086
7
49
87

Given the fact that I have close to 120 files and each spans over 3-10 different days, not sure adding points to them will work, but I understand your idea, sadly I think that it would be too tedious to do to each file – Richard Michaud Langis Jun 13 '22 at 14:11
Unless there is a way to add a line to each day at 8am and 5pm, but not sure that is possible – Richard Michaud Langis Jun 13 '22 at 14:13
@RichardMichaudLangis that's exactly what my code does - it takes the dates from the data frame and automatically inserts invisible points at 8am and 5pm on each unique day. Try it. – Allan Cameron Jun 13 '22 at 14:18
@RichardMichaudLangis there only _is_ one. The bottom one is just the data I had to make up because your sample data only has one date in it - you don't need that. This is just a dummy version of your actual data frame. – Allan Cameron Jun 13 '22 at 14:29
So I tried your graphing code up top, this seems to work in terms of forcing the limits of my graph but does not keep the right limits, it starts at 2 am and ends at 5pm – Richard Michaud Langis Jun 13 '22 at 14:29
I will post the output into my original question – Richard Michaud Langis Jun 13 '22 at 14:30
@RichardMichaudLangis that's not my graphing code that you have posted. That's your graphing code. – Allan Cameron Jun 13 '22 at 14:32
Apologies, here you go should be yours now – Richard Michaud Langis Jun 13 '22 at 14:35
1

@RichardMichaudLangis this seems to be a time zone issue. Try changing `rep(c(8, 17)` in the above code to `rep(c(12, 21)` and the bit that says `rep(c(8, 11, 14, 17)` to `rep(c(12, 15, 18, 21)` (the times all need 4 hours added on) – Allan Cameron Jun 13 '22 at 14:45
EUREKA ! Thank you very much, been working on this for a a few days, Works great now – Richard Michaud Langis Jun 13 '22 at 14:47
1

what I always find admirable about you that you understand the actual problem of the OP despite sometimes unclear way to ask.. Also your willingness to create data that suits the problem better. +1 – tjebo Jun 13 '22 at 16:49

tjebo · Answer 2 · 2022-06-13T17:10:31.707

If you are faceting by day anyways, and only showing time on the x axis, then the time variable does not need to be in a datetime class. You can just convert it to continuous data first, and then the whole shebang becomes really simple.

Using Allan's data. (thanks)

library(tidyverse)
set.seed(1)

graph_POST <- data.frame(Date = rep(as.Date(
  c("2022-06-02", "2022-06-03", "2022-06-06", "2022-06-07", "2022-06-08")),
  each = 100),
  Time = as.POSIXct(rep(as.Date(
    c("2022-06-02", "2022-06-03", "2022-06-06", "2022-06-07", "2022-06-08")),
    each = 100)) + sample(32000:61200, 500),
  Standing = sample(c(0, 4), 500, TRUE, prob = c(4, 1)),
  Stepping = sample(c(0, 3), 500, TRUE, prob = c(4, 1)),
  Cycling = sample(c(0, 2), 500, TRUE, prob = c(4, 1)),
  New_Sitting = sample(0:1, 500, TRUE, prob = c(4, 1)))

graph_POST %>%
  mutate(day = lubridate::day(Date), 
         hour = as.integer(format(Time,"%H")), 
## If you need it precise to the minute
         minute = as.integer(format(Time,"%M")) ,
         hour_minute = hour + minute/60) %>%
  pivot_longer(Standing:New_Sitting, names_to = "Posture") %>%
  ggplot(aes(x = hour_minute, y = value, color = Posture)) +
  geom_point() +
  facet_wrap(~day) +
  ## now setting the limits is easy.
  scale_x_continuous(breaks= 8:17, limits = c(8, 17)) 
#> Warning: Removed 252 rows containing missing values (geom_point).

Very nice @tjebo, though you lose a little detail with `hour` rounding down. Adding minutes / 60 would fix this without ruining the overall idea though. — Allan Cameron, Jun 13 '22 at 16:56
@AllanCameron true that - I have amended the answer slightly. — tjebo, Jun 13 '22 at 17:03

Forcing the Time Axes to be set at a specific boundaries

2 Answers2