Lubridate and ggplot

1 Load packages

Show the code 
library(tidyverse)
library(nycflights13)
library(ggh4x)
library(ggforce)
1
A collection of packages including the lubridate and ggplot2 packagse.
2
A package that includes the flights and airlines data used in this tutorial.

The nycflights13 package comes with a number of data sets including flights and airlines that will be used in this exercise.

2 Wrangle data

Show the code 
flights <- flights %>% 
select(carrier, time_hour, arr_delay)

flights %>% 
    head(10)
1
The carrier variable provides an abbreviation of the name of the airline, the time_hour variable provides the date and time of the flight and the arr_delay provides the time in minutes that the flight arrived late or early. 
# A tibble: 10 × 3
   carrier time_hour           arr_delay
   <chr>   <dttm>                  <dbl>
 1 UA      2013-01-01 05:00:00        11
 2 UA      2013-01-01 05:00:00        20
 3 AA      2013-01-01 05:00:00        33
 4 B6      2013-01-01 05:00:00       -18
 5 DL      2013-01-01 06:00:00       -25
 6 UA      2013-01-01 05:00:00        12
 7 B6      2013-01-01 06:00:00        19
 8 EV      2013-01-01 06:00:00       -14
 9 B6      2013-01-01 06:00:00        -8
10 AA      2013-01-01 06:00:00         8
Show the code 
airlines %>%
    head(10)
1
Notice that the “carrier” variable is the same as is found in the “flights” dataset. This fact will allow us to join the two data sets. 
# A tibble: 10 × 2
   carrier name                       
   <chr>   <chr>                      
 1 9E      Endeavor Air Inc.          
 2 AA      American Airlines Inc.     
 3 AS      Alaska Airlines Inc.       
 4 B6      JetBlue Airways            
 5 DL      Delta Air Lines Inc.       
 6 EV      ExpressJet Airlines Inc.   
 7 F9      Frontier Airlines Inc.     
 8 FL      AirTran Airways Corporation
 9 HA      Hawaiian Airlines Inc.     
10 MQ      Envoy Air
Show the code 
flights <- flights %>% 
   left_join(airlines, by = "carrier")

flights %>% 
    head(10)
1
The left_join() function merges two data frames based on one or more common columns (keys). The resulting data frame includes all rows from the left data frame (first argument) and matching rows from the right data frame (second argument). In this case, because flights is being piped into the function, it is the first argument by default. 
# A tibble: 10 × 4
   carrier time_hour           arr_delay name                    
   <chr>   <dttm>                  <dbl> <chr>                   
 1 UA      2013-01-01 05:00:00        11 United Air Lines Inc.   
 2 UA      2013-01-01 05:00:00        20 United Air Lines Inc.   
 3 AA      2013-01-01 05:00:00        33 American Airlines Inc.  
 4 B6      2013-01-01 05:00:00       -18 JetBlue Airways         
 5 DL      2013-01-01 06:00:00       -25 Delta Air Lines Inc.    
 6 UA      2013-01-01 05:00:00        12 United Air Lines Inc.   
 7 B6      2013-01-01 06:00:00        19 JetBlue Airways         
 8 EV      2013-01-01 06:00:00       -14 ExpressJet Airlines Inc.
 9 B6      2013-01-01 06:00:00        -8 JetBlue Airways         
10 AA      2013-01-01 06:00:00         8 American Airlines Inc.
Show the code 
flights <- flights %>% 
    filter(name %in% c(
        "American Airlines Inc.",
        "Delta Air Lines Inc.",
        "JetBlue Airways",
        "United Air Lines Inc."))

flights %>% 
    head(10)
1
using %in% when filtering allows you to create a list of possible values in the variable being filtered. An observation with any of the listed values will be kept in the filtered data set. 
# A tibble: 10 × 4
   carrier time_hour           arr_delay name                  
   <chr>   <dttm>                  <dbl> <chr>                 
 1 UA      2013-01-01 05:00:00        11 United Air Lines Inc. 
 2 UA      2013-01-01 05:00:00        20 United Air Lines Inc. 
 3 AA      2013-01-01 05:00:00        33 American Airlines Inc.
 4 B6      2013-01-01 05:00:00       -18 JetBlue Airways       
 5 DL      2013-01-01 06:00:00       -25 Delta Air Lines Inc.  
 6 UA      2013-01-01 05:00:00        12 United Air Lines Inc. 
 7 B6      2013-01-01 06:00:00        19 JetBlue Airways       
 8 B6      2013-01-01 06:00:00        -8 JetBlue Airways       
 9 AA      2013-01-01 06:00:00         8 American Airlines Inc.
10 B6      2013-01-01 06:00:00        -2 JetBlue Airways
Show the code 
flights <- flights %>%
  mutate(month = month(time_hour, label = TRUE, abbr = FALSE))

flights %>% 
    head(10)
1
Here the mutate function is creating a variable called month using the month function from lublidate. Label = TRUE ensure that the actual name of the month and not a number is used. 
# A tibble: 10 × 5
   carrier time_hour           arr_delay name                   month  
   <chr>   <dttm>                  <dbl> <chr>                  <ord>  
 1 UA      2013-01-01 05:00:00        11 United Air Lines Inc.  January
 2 UA      2013-01-01 05:00:00        20 United Air Lines Inc.  January
 3 AA      2013-01-01 05:00:00        33 American Airlines Inc. January
 4 B6      2013-01-01 05:00:00       -18 JetBlue Airways        January
 5 DL      2013-01-01 06:00:00       -25 Delta Air Lines Inc.   January
 6 UA      2013-01-01 05:00:00        12 United Air Lines Inc.  January
 7 B6      2013-01-01 06:00:00        19 JetBlue Airways        January
 8 B6      2013-01-01 06:00:00        -8 JetBlue Airways        January
 9 AA      2013-01-01 06:00:00         8 American Airlines Inc. January
10 B6      2013-01-01 06:00:00        -2 JetBlue Airways        January
Show the code 
delay_monthly <- flights %>%
    group_by(name, month) %>%
  summarise(avg_arr_delay = mean(arr_delay, na.rm = TRUE))

delay_monthly %>% 
    head()
1
We group by both name and month because we want to have the average delay of flights for every month of every selected airline.
2
The summarise function will create a variable, in this case called “avg_arr_delay” which will have values for the mean arr_delay for each month of each airline (see group_by() above). 
# A tibble: 6 × 3
# Groups:   name [1]
  name                   month    avg_arr_delay
  <chr>                  <ord>            <dbl>
1 American Airlines Inc. January          0.982
2 American Airlines Inc. February         1.09 
3 American Airlines Inc. March           -2.98 
4 American Airlines Inc. April            8.26 
5 American Airlines Inc. May             -4.68 
6 American Airlines Inc. June             6.48
Show the code 
delay_monthly <- delay_monthly %>% 
  mutate(delay_color = ifelse(avg_arr_delay < 0, "darkgreen", "darkred"))

delay_monthly %>% 
    head(10)
1
The ifelse() function here states that if the avg_arr_delay value is less than zero, then the new variable being created (delay_color) will have the value “darkgreen”. If not, the it will have the value “darkred”. 
# A tibble: 10 × 4
# Groups:   name [1]
   name                   month     avg_arr_delay delay_color
   <chr>                  <ord>             <dbl> <chr>      
 1 American Airlines Inc. January           0.982 darkred    
 2 American Airlines Inc. February          1.09  darkred    
 3 American Airlines Inc. March            -2.98  darkgreen  
 4 American Airlines Inc. April             8.26  darkred    
 5 American Airlines Inc. May              -4.68  darkgreen  
 6 American Airlines Inc. June              6.48  darkred    
 7 American Airlines Inc. July              4.32  darkred    
 8 American Airlines Inc. August           -2.30  darkgreen  
 9 American Airlines Inc. September        -8.57  darkgreen  
10 American Airlines Inc. October          -3.79  darkgreen

3 Data visualization

Show the code 
delay_monthly %>% 
    ggplot(aes(x = month, y = avg_arr_delay)) +
  geom_col()

Show the code 
ggplot(delay_monthly, aes(x = month, y = avg_arr_delay)) +
  geom_col() +
  facet_wrap(~ name)

Show the code 
ggplot(delay_monthly, aes(x = month, y = avg_arr_delay, fill = delay_color)) +
  geom_col(alpha = 0.6) +
  facet_wrap(~ name)+
  scale_fill_identity() +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 45,
                                   hjust = 1))
1
This essentially says “use the values in the fill aesthetic exactly as they are, they’re already colors.”
2
This is a good theme for faceted plots as it creates a black border around each facet.
3
angles the axis text to 45 degrees so that the month names don’t overlap.
4
With the text now angled, R needs to know where to place the text so that it doesn’t overlap with the axis line.

Show the code 
ggplot(delay_monthly, aes(x = month, y = avg_arr_delay, fill = delay_color)) +
  geom_col(alpha= 0.6) +
  facet_wrap(~ name)+
  labs(title = "Arrival Delay per Month by  Carrier (2013)",
       x = "",
       y = "Average Arrival Delay (minutes)") +
  scale_fill_identity() +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))
1
Note that I have not included a lable for the X axis as it is clear that its months

Show the code 
annotation_text <- data.frame(
  x = 8,
  y = 15,
  name = "American Airlines Inc.")

annotation_arrow <- data.frame(
  x = 9.1,
  y = 11,
  xend = 9,
  yend = 0.2,
  name = "American Airlines Inc.")

ggplot(delay_monthly, aes(x = month, y = avg_arr_delay, fill = delay_color)) +
  geom_col(alpha= 0.6) +
  facet_wrap(~ name)+
  labs(title = "Arrival Delay per Month by  Carrier (2013)", 
       x = "", 
       y = "Average Arrival Delay (minutes)") + 
  scale_fill_identity() +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))+
    geom_text(data = annotation_text,
            aes(x = x,
                y = y,
                label = "Good month\n to travel"),
                color = "darkgreen",
                fontface = "bold",
                size = 3.5,
                inherit.aes = FALSE) +
  geom_curve(data = annotation_arrow,
             aes(x = x,
                 y = y,
                 xend = xend,
                 yend = yend),
                 arrow = arrow(length = unit(0.33, "cm")),
                 color = "darkgreen",
                 curvature = -0.25,
                 linewidth = 1.2,
                 inherit.aes = FALSE)
1
This data frame contains the x and y coordinates for the text annotation as well as the name of the facet into which the annotation will be placed.
2
This data frame contains the x and y coordinates for beginning and end of the arrow.
3
We point ggplot to the data frame with the correct coordinates.
4
We indicate that for the aesthetics, the geom_text() function is to use the x and y values from the data frame.
5
We tell geom_text() to create a new line after the word month.
6
We indicate further parameters of the text.
7
By default, geom_text() would inherit the aesthetic mappings (like x, y, fill) defined in the main ggplot() call. Here we prevents this and tells geom_text() to only use the aesthetics defined within its own aes() function and the data argument.
8
We tell geom_curve() to use the data from the data frame created above.
9
We define the parameters of the curve.
10
We indicate that geom_curve is to use its own aes() function and the data argument.

4 The lubridate package

The lubridate package in R simplifies working with date-times by providing intuitive and user-friendly functions. It is part of the tidyverse ecosystem and makes it easy to parse, manipulate, extract, and perform arithmetic operations on date-time data.

4.1 Lubridate primarily helps you:

  • Parse character or numeric date-time data into R date-time formats.

  • Manipulate dates and times through arithmetic and transformations.

  • Extract components like years, months, days, hours, minutes, and seconds.

  • Handle time zones effectively.

  • Perform interval and duration calculations seamlessly.

    4.2 Key Functions in Lubridate:

4.3 Parsing Functions:

  • ymd(), mdy(), dmy(), ydm(), myd(), dym() Parse character or numeric vectors into date objects, based on the order of year, month, and day.
  • ymd_hms(), ymd_hm(), mdy_hms(), dmy_hms(), etc. Parse date-time strings with specified order, returning POSIXct date-time objects.

4.4 Date-time Component Extraction:

  • year() Extracts the year from a date or date-time.
  • month() Extracts the month (numeric by default) from a date or date-time.
  • day() Extracts the day of the month.
  • wday() Extracts the weekday (numeric or label).
  • qday() Extracts the day of the quarter.
  • yday() Extracts the day of the year.
  • hour(), minute(), second() Extract hour, minute, or second from a date-time object.

4.5 Date-time Modification:

  • update() Changes individual components (year, month, day, hour, etc.) of a date-time object.
  • floor_date(), ceiling_date(), round_date() Round a date-time to the nearest specified unit (e.g., day, month, year).

4.6 Time Span Classes:

  • duration() Creates duration objects (exact number of seconds).
  • period() Creates period objects (human-readable units such as months and years).
  • interval() Defines intervals between two date-time points.

4.7 Arithmetic and Interval Operations:

  • %–% Creates an interval between two dates or times (e.g., start_date %–% end_date).
  • int_length() Computes the length of an interval in seconds.
  • int_start(), int_end() Extract start or end dates from intervals.
  • int_shift() Shifts an interval by a specified duration or period.
  • int_overlaps() Checks if intervals overlap.

4.8 Time Zones:

  • with_tz() Converts the time to a different time zone without changing clock time.
  • force_tz() Forces the current time to a new time zone, keeping the clock time fixed.
  • tz() Retrieves or assigns the time zone attribute of a date-time object.

4.9 Logical Checks and Helpers:

  • is.Date(), is.POSIXct(), is.POSIXlt() Check if an object is of the respective date-time class.
  • leap_year() Tests whether a given year is a leap year.
  • days_in_month() Returns the number of days in a given month.

4.10 Convenience Functions:

  • today() Returns the current date.
  • now() Returns the current date and time.
  • make_date(), make_datetime() Construct date or date-time objects directly from components.
  • stamp() Creates custom date-time formatting functions based on an example format.
  • pretty_dates() Generates visually appealing sequences of dates for plotting axes.