How to plot hourly temperature data group by week in ggplot2?

Question

I have weather data collected in one-week-long experiments, with each week represented by the first day of the experiment, as shown below.

I want to create a graph to visualize hourly weather data within each week, where each column/line represents one hour's data for each week. To make x-axis similar to my experimental data figure, I added a date2 column using mutate(date2 = first(date)). I had to create date2 because the duration of the experiment was not exactly one week, I can't use 1 week in scale_x_date()

I want a create a plot that resembles the following graph, but I'd like to plot hourly data rather than daily data. My data is already collected at hourly intervals.

My weather data looks like this

Here is the reproducible example

df <-
  structure(
    list(
      date = c(
        "05/26/2014 11:00 AM",
        "05/26/2014 12:00 PM",
        "05/26/2014 01:00 PM",
        "06/09/2014 12:00 PM",
        "06/09/2014 01:00 PM",
        "06/09/2014 02:00 PM",
        "06/16/2014 12:00 PM",
        "06/16/2014 01:00 PM",
        "06/16/2014 02:00 PM",
        "06/23/2014 12:00 PM",
        "06/23/2014 01:00 PM",
        "06/23/2014 02:00 PM",
        "06/30/2014 12:00 PM",
        "06/30/2014 01:00 PM",
        "06/30/2014 02:00 PM",
        "07/07/2014 12:00 PM",
        "07/07/2014 01:00 PM",
        "07/07/2014 02:00 PM",
        "07/14/2014 12:00 PM",
        "07/14/2014 01:00 PM",
        "07/14/2014 02:00 PM",
        "07/21/2014 12:00 PM",
        "07/21/2014 01:00 PM",
        "07/21/2014 02:00 PM",
        "07/28/2014 12:00 PM",
        "07/28/2014 01:00 PM",
        "07/28/2014 02:00 PM",
        "08/04/2014 12:00 PM",
        "08/04/2014 01:00 PM",
        "08/04/2014 02:00 PM",
        "08/11/2014 12:00 PM",
        "08/11/2014 01:00 PM",
        "08/11/2014 02:00 PM",
        "09/08/2014 12:00 PM",
        "09/08/2014 01:00 PM",
        "09/08/2014 02:00 PM",
        "09/15/2014 12:00 PM",
        "09/15/2014 01:00 PM",
        "09/15/2014 02:00 PM",
        "09/22/2014 12:00 PM",
        "09/22/2014 01:00 PM",
        "09/22/2014 02:00 PM",
        "09/29/2014 12:00 PM",
        "09/29/2014 01:00 PM",
        "09/29/2014 02:00 PM",
        "10/06/2014 12:00 PM",
        "10/06/2014 01:00 PM",
        "10/06/2014 02:00 PM",
        "10/13/2014 12:00 PM",
        "10/13/2014 01:00 PM",
        "10/13/2014 02:00 PM",
        "10/20/2014 12:00 PM",
        "10/20/2014 01:00 PM",
        "10/20/2014 02:00 PM",
        "10/27/2014 12:00 PM",
        "10/27/2014 01:00 PM",
        "10/27/2014 02:00 PM",
        "11/03/2014 12:00 PM",
        "11/03/2014 01:00 PM",
        "11/03/2014 02:00 PM"
      ),
      date2 = structure(
        c(
          1401104700,
          1401104700,
          1401104700,
          1402317900,
          1402317900,
          1402317900,
          1402920000,
          1402920000,
          1402920000,
          1403524800,
          1403524800,
          1403524800,
          1404129600,
          1404129600,
          1404129600,
          1404734400,
          1404734400,
          1404734400,
          1405339200,
          1405339200,
          1405339200,
          1405944000,
          1405944000,
          1405944000,
          1406548800,
          1406548800,
          1406548800,
          1407153600,
          1407153600,
          1407153600,
          1407758400,
          1407758400,
          1407758400,
          1410177600,
          1410177600,
          1410177600,
          1410782400,
          1410782400,
          1410782400,
          1411387200,
          1411387200,
          1411387200,
          1411992000,
          1411992000,
          1411992000,
          1412596800,
          1412596800,
          1412596800,
          1413201600,
          1413201600,
          1413201600,
          1413806400,
          1413806400,
          1413806400,
          1414411200,
          1414411200,
          1414411200,
          1415016000,
          1415016000,
          1415016000
        ),
        tzone = "UTC",
        class = c("POSIXct",
                  "POSIXt")
      ),
      week = structure(
        c(
          1L,
          1L,
          1L,
          2L,
          2L,
          2L,
          3L,
          3L,
          3L,
          4L,
          4L,
          4L,
          5L,
          5L,
          5L,
          6L,
          6L,
          6L,
          7L,
          7L,
          7L,
          8L,
          8L,
          8L,
          9L,
          9L,
          9L,
          10L,
          10L,
          10L,
          11L,
          11L,
          11L,
          12L,
          12L,
          12L,
          13L,
          13L,
          13L,
          14L,
          14L,
          14L,
          15L,
          15L,
          15L,
          16L,
          16L,
          16L,
          17L,
          17L,
          17L,
          18L,
          18L,
          18L,
          19L,
          19L,
          19L,
          20L,
          20L,
          20L
        ),
        levels = c(
          "1",
          "2",
          "3",
          "4",
          "5",
          "6",
          "7",
          "8",
          "9",
          "10",
          "11",
          "12",
          "13",
          "14",
          "15",
          "16",
          "17",
          "18",
          "19",
          "20",
          "21",
          "22",
          "23",
          "24",
          "25",
          "26"
        ),
        class = "factor"
      ),
      mean_temperature = c(
        25,
        25.25,
        26.075,
        26.2,
        26.35,
        27.65,
        29.95,
        29.4,
        30.5,
        22.75,
        24,
        25.275,
        25.6,
        26.8,
        27.125,
        27.5,
        27.525,
        26.1,
        30.65,
        30.65,
        30.425,
        26.15,
        24.5,
        23.7,
        25.65,
        25.85,
        25.2,
        26.975,
        27.125,
        28.275,
        18.55,
        18.775,
        19.125,
        18.8,
        20,
        19.375,
        17.825,
        17.925,
        18.25,
        18.975,
        20.825,
        20.1,
        18.1,
        18.4,
        19.2,
        16.325,
        18.4,
        18.75,
        14.675,
        14.925,
        15.3,
        15.9,
        17.175,
        17.525,
        22.45,
        23,
        23.475,
        17.8,
        18.05,
        17.725
      )
    ),
    class = c("grouped_df", "tbl_df", "tbl",
              "data.frame"),
    row.names = c(NA,-60L),
    groups = structure(
      list(
        week = structure(
          1:20,
          levels = c(
            "1",
            "2",
            "3",
            "4",
            "5",
            "6",
            "7",
            "8",
            "9",
            "10",
            "11",
            "12",
            "13",
            "14",
            "15",
            "16",
            "17",
            "18",
            "19",
            "20",
            "21",
            "22",
            "23",
            "24",
            "25",
            "26"
          ),
          class = "factor"
        ),
        .rows = structure(
          list(
            1:3,
            4:6,
            7:9,
            10:12,
            13:15,
            16:18,
            19:21,
            22:24,
            25:27,
            28:30,
            31:33,
            34:36,
            37:39,
            40:42,
            43:45,
            46:48,
            49:51,
            52:54,
            55:57,
            58:60
          ),
          ptype = integer(0),
          class = c("vctrs_list_of",
                    "vctrs_vctr", "list")
        )
      ),
      row.names = c(NA,-20L),
      .drop = TRUE,
      class = c("tbl_df",
                "tbl", "data.frame")
    )
  )

EDIT

Column graph are not appropriate to plot hourly temperature data because every data point starts from zero, and I don't have zeros in my data. Appropriate graph would be point or line graph as shown here or here

The data is not properly copied and pasted, I'm getting an error when trying to reproduce it in my R session, `Error: C stack usage 15927504 is too close to the limit`. — Rui Barradas, Jul 30 '23 at 17:39
@RuiBarradas My apologies. I have copied the data again. Should be okay now. — Ahsk, Jul 30 '23 at 17:47

Rui Barradas · Answer 1 · 2023-07-31T07:08:27.093

Like this?
Extract hour from date, coerce the former to factor and the latter to a date object. Then plot grouping by hour. The y axis breaks by week.

Note that I process the date one step at a time in order to make it more readable and easier to debug, if needed.

suppressPackageStartupMessages({
  library(dplyr)
  library(lubridate)
  library(ggplot2)
})

df %>% 
  ungroup() %>%
  mutate(
    date = mdy_hm(date),
    hour = hour(date),
    date = ymd(format(date, "%Y-%m-%d")),
    hour = factor(hour)
  ) %>% 
  ggplot(aes(date, mean_temperature, group = hour)) +
  geom_col(position = position_dodge(), fill = "steelblue", color = "white") +
  scale_x_date(date_breaks = "1 week") +
  coord_flip() +
  theme_bw()

^{Created on 2023-07-30 with reprex v2.0.2}

Edit

To remove the days without measurements, coerce the x-axis variable, date, to factor and remove the x-axis date scale.

Note that I have kept the white line separation between bars, to remove it remove color = "white" from geom_col.

suppressPackageStartupMessages({
  library(dplyr)
  library(lubridate)
  library(ggplot2)
})

df %>% 
  ungroup() %>%
  mutate(
    date = mdy_hm(date),
    hour = hour(date),
    date = ymd(format(date, "%Y-%m-%d")),
    date = factor(date),
    hour = factor(hour)
  ) %>% 
  ggplot(aes(date, mean_temperature, group = hour)) +
  geom_col(position = position_dodge(), fill = "steelblue", color = "white") +
  labs(x = "Date", y = "Mean temperature") +
  # coord_flip() +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1))

^{Created on 2023-07-31 with reprex v2.0.2}

@Ahsk I don't understand, you are saying that the bars shouldn't start at 0? — Rui Barradas, Aug 03 '23 at 21:38
Yes, the bars should correspond to the actual range of temperature in that hour. If the starts bars at zero then that would mean zero temperature in that hour, right? .The data is recorded at hourly interval and there is no zero is the `mean_temperature` column, but the bars are starting at zero. — Ahsk, Aug 03 '23 at 21:42
@Ahsk No, I believe you are wrong, the bars should start at zero. What can be misleading is to have a bar plot starting at a fluid point, in this case the data minimum. Though this is not the case, it would make it easier to manipulate the plots, have the origin change and the bars will look smaller or bigger at will. — Rui Barradas, Aug 03 '23 at 21:57

How to plot hourly temperature data group by week in ggplot2?

1 Answers1

Edit