Time-series Boxplot using precomputed quantiles with ggplot2

Question

If I run a time series box plot on an irregularly spaced time series (with missing values) with the date field as chr I get a box plot but with no gaps where the data is missing. If I import the data field as date and try the same plot I get the following error message:

Error in geom_boxplot(): ! Problem while converting geom to grob. ℹ Error occurred in the 1st layer. Caused by error in draw_group(): ! Can only draw one boxplot per group ℹ Did you forget aes(group = ...)?

I am not sure what variable I should be grouping on to get this to work. A sample of my data:

# A tibble: 17 × 9
   SmplDate   minDPM q1DPM medianDPM q3DPM maxDPM meanDPM stdevsDPM MeanBiomass
   <date>      <dbl> <dbl>     <dbl> <dbl>  <dbl>   <dbl>     <dbl>       <dbl>
 1 2023-05-30    3    4.5       5.5   7.5    23.5    6.74      3.74        2.85
 2 2005-06-02    6   16.4      20.5  25.5    52     21.9       9.87        7.56
 3 2006-06-09    4    6.5      10.5  13.9    22.5   10.5       4.51        4.32
 4 2007-06-07    4    7.5      10.5  13.9    36     11.5       5.67        4.65
 5 2008-06-02    0.5  4         5     7      15.5    5.66      2.73        2.36
 6 2009-06-01    2    6         8.25 11.9    23      9.39      4.23        3.91
 7 2010-06-01    2.5  7.12     11    15.4    26.5   11.8       5.80        4.73
 8 2011-05-30    1.5  4         5.25  8      18      6.6       3.88        2.79
 9 2012-06-01    1.5  2.62      3.5   4.5     7.5    3.72      1.33        1.34
10 2013-06-05    2.5  5.12      6.25  8.5    17      7.01      2.73        2.96
11 2014-06-07    1.5  3         4     6.5    11.5    4.7       2.34        1.88
12 2016-05-31    3.5  8.25     12.5  16      59     13.3       8.11        5.21
13 2017-06-05    2    3.5       4     5      11.5    4.5       1.81        1.78
14 2018-06-08    1    4.62      6.25  9.88   28      7.78      5.09        3.29
15 2019-05-31    3    5.5       7     9      25      8.03      3.92        3.39
16 2021-06-05    3    6.5       8    12      24.5    9.41      4.42        3.91
17 2022-06-02    2    4.5       5.5   7.5    13      6.23      2.27        2.62

MY code so far:

# Time series boxplot
ggplot(data = vca01,
       aes(x = SmplDate)) +
    geom_boxplot(
        aes(
           ymin = minDPM,
           lower = q1DPM,
           middle = medianDPM,
           upper = q3DPM,
           ymax = maxDPM
           ),
          stat = "identity") +
  theme_bw() +
  labs(x = "Year",
       y = "Grass Biomass",
       title = "Site name: VCA01")

I am wondering if this can be achieved if I converted the tibble to a zoo object? This is what I tried:

#convert tibble to zoo object
tseries <- read.zoo(vca01)
class(tseries)

# Time series Boxplot
ggplot(data = tseries,
        aes(x = Index,
            ymin = minDPM,
            lower = q1DPM,
            middle = medianDPM,
            upper = q3DPM,
            ymax = maxDPM)) +
  geom_boxplot(stat = "identity") +
  theme_bw() +
  labs(x = "Year",
       y = "Grass Biomass",
       title = "Site name: VCA01")`

Error message: Error in geom_boxplot(): ! Problem while converting geom to grob. ℹ Error occurred in the 1st layer. Caused by error in draw_group(): ! Can only draw one boxplot per group ℹ Did you forget aes(group = ...)?

I tried the ggplot code with the data field as date and got the error described. I am not sure that melt (as suggested in R: Plot a time series with quantiles using ggplot2) will make any difference as my data are already in long format.

score 0 · Answer 1 · answered Jun 06 '23 at 12:01

As SmplDate is a continuous variable you have to be more explicit about the grouping in case of geom_boxplot, i.e. you have to map SmplDate on the group aes:

library(ggplot2)

ggplot(
  data = vca01,
  aes(x = SmplDate)
) +
  geom_boxplot(
    aes(
      ymin = minDPM,
      lower = q1DPM,
      middle = medianDPM,
      upper = q3DPM,
      ymax = maxDPM,
      group = SmplDate
    ),
    stat = "identity"
  ) +
  theme_bw() +
  labs(
    x = "Year",
    y = "Grass Biomass",
    title = "Site name: VCA01"
  )

DATA

structure(list(SmplDate = structure(c(19507, 12936, 13308, 13671, 
14032, 14396, 14761, 15124, 15492, 15861, 16228, 16952, 17322, 
17690, 18047, 18783, 19145), class = "Date"), minDPM = c(3, 6, 
4, 4, 0.5, 2, 2.5, 1.5, 1.5, 2.5, 1.5, 3.5, 2, 1, 3, 3, 2), q1DPM = c(4.5, 
16.4, 6.5, 7.5, 4, 6, 7.12, 4, 2.62, 5.12, 3, 8.25, 3.5, 4.62, 
5.5, 6.5, 4.5), medianDPM = c(5.5, 20.5, 10.5, 10.5, 5, 8.25, 
11, 5.25, 3.5, 6.25, 4, 12.5, 4, 6.25, 7, 8, 5.5), q3DPM = c(7.5, 
25.5, 13.9, 13.9, 7, 11.9, 15.4, 8, 4.5, 8.5, 6.5, 16, 5, 9.88, 
9, 12, 7.5), maxDPM = c(23.5, 52, 22.5, 36, 15.5, 23, 26.5, 18, 
7.5, 17, 11.5, 59, 11.5, 28, 25, 24.5, 13), meanDPM = c(6.74, 
21.9, 10.5, 11.5, 5.66, 9.39, 11.8, 6.6, 3.72, 7.01, 4.7, 13.3, 
4.5, 7.78, 8.03, 9.41, 6.23), stdevsDPM = c(3.74, 9.87, 4.51, 
5.67, 2.73, 4.23, 5.8, 3.88, 1.33, 2.73, 2.34, 8.11, 1.81, 5.09, 
3.92, 4.42, 2.27), MeanBiomass = c(2.85, 7.56, 4.32, 4.65, 2.36, 
3.91, 4.73, 2.79, 1.34, 2.96, 1.88, 5.21, 1.78, 3.29, 3.39, 3.91, 
2.62)), row.names = c("1", "2", "3", "4", "5", "6", "7", "8", 
"9", "10", "11", "12", "13", "14", "15", "16", "17"), class = "data.frame")

Time-series Boxplot using precomputed quantiles with ggplot2

1 Answers1