Tidy Tuesday: Daylight Savings Time

R
TidyTuesday
R-code
code-along
Data-Viz
ggimage
maps
TidyTuesday: Exploring Time Zones and Daylight Savings Time
Author

Louise E. Sinks

Published

March 28, 2023

Modified

November 3, 2023

This week’s TidyTuesday is about the timezone data from IANA timezone database.

library(tidytuesdayR)
library(tidyverse)
library(skimr)
library(ggthemes)
library(gt)
library(lubridate)
library(skimr)
library(lutz)
library(maps)
library(scales)
library(sf)
library(ggimage)

The history of this database is fascinating. It is used by many computer systems to determine the correct time based on location. To learn more, I recommend reading Daniel Rosehill’s article on the topic. For a drier history, check out the wikipedia article.

# Get the Data

# Read in with tidytuesdayR package 
# This loads the readme and all the datasets for the week of interest

# Either ISO-8601 date or year/week works!

#tuesdata <- tidytuesdayR::tt_load('2023-03-21')
tuesdata <- tidytuesdayR::tt_load(2023, week = 13)

transitions <- tuesdata$transitions
timezones <- tuesdata$timezones
timezone_countries <- tuesdata$timezone_countries
countries <- tuesdata$countries

It is suggested that we change the begin and end variables in transitions to datetimes.

transitions <- transitions %>%
  mutate(begin = as_datetime(begin), end = as_datetime(end))

I was interested in how many countries had multiple times zones. I know the US has 4 time zones in the continental US.

num_zones <- timezone_countries %>%
  count(country_code, sort = TRUE)

num_zones %>% 
  filter(n > 1) %>%
  left_join(countries) %>%
  select(place_name, n) %>%
  filter(place_name != "NA") %>%
  gt() %>%
  cols_label(place_name = "Country", n = "Number of TZs") %>%
  opt_stylize(style = 6, color = "blue", add_row_striping = TRUE) %>%
  tab_header(title = "Countries with Multiple TZs")
Countries with Multiple TZs
Country Number of TZs
United States 29
Canada 28
Russia 27
Brazil 16
Argentina 12
Australia 12
Mexico 11
Kazakhstan 7
Greenland 4
Indonesia 4
Ukraine 4
Chile 3
Spain 3
Micronesia 3
Kiribati 3
Mongolia 3
Malaysia 3
French Polynesia 3
Portugal 3
US minor outlying islands 3
Congo (Dem. Rep.) 2
China 2
Cyprus 2
Germany 2
Ecuador 2
Marshall Islands 2
New Zealand 2
Papua New Guinea 2
Palestine 2
French Southern & Antarctic Lands 2
Uzbekistan 2
Vietnam 2

And we find that the United States has 29!! time zones in the database. This was unexpected, so say the least. I thought maybe there were some times zones for territories and perhaps military bases that I did not know about. I also thought there might be some extra time zones arising from some states using daylight savings time, while others in the same area might not. I wanted to visualize where these times zones were.

US_tz <- timezone_countries %>% 
  filter(country_code == "US") %>%
  left_join(timezones)
Joining with `by = join_by(zone)`

I found the lutz package created nice pictograms about when a timezone shifts from DST and back. (This package uses the same underlying database that we are using here to determine when the shifts occur.)

 tz_plot(US_tz$zone[21])

I created the plots and saved them as images. I modified a function I found on stack overflow to create the file names.

wd <- getwd()
filepath = file.path(wd)


make_filename = function(number){
  # doing this, putting it all on a single line or using pipe %>%
  # is just matter of style
  filename = paste("tzplot", number, sep="_")
  filename = paste0(filename, ".png")
  filename = file.path(filepath, filename)
  
  filename
}

#creating a variable to store the files name
US_tz <- US_tz %>%
  mutate(image_name = "tbd")

index <- 1
for (index in seq(1, nrow(US_tz))) {
  filename = make_filename(index)
  US_tz[index , "image_name"] <- filename
  # 1. Open jpeg file
  png(filename, width = 350, height = 350, bg = "transparent")
  # 2. Create the plot
  # you need to print the plot if you call it inside a loop
  print(tz_plot(US_tz$zone[index]))
  # 3. Close the file
  dev.off()
  index = index + 1
}

Next I created a world map, inspired by the one from

I hadn’t previously used the maps package, so I appreciate being introduced to it. The maps package only has a mainland US map, so I used the world map. (Plus, as I mentioned, I thought some of these time zones would be in other parts of the world.) I followed a tutorial on Plotting Points as Images in ggplot and used the hints about aspect ratio to make my tz_plot circles remain circular. However, that did stretch the world a bit.

aspect_ratio <- 1.618  

us_tz_map <- map_data("world") %>% 
  ggplot(aes(long, lat)) +
  geom_polygon(aes(group = group), fill = "white", 
               color = "gray30", alpha = 0.9) +
  geom_image(aes(x = longitude, latitude, image = image_name), 
             data = US_tz, size = 0.025, by = "width",
             asp = aspect_ratio) +
  coord_sf() +
  labs(title = "The United States has 29 Timezone- Mostly Redunant",
       caption = "Data from: https://data.iana.org/time-zones/tz-link.html") +
  theme_void() +
  theme(aspect.ratio = 1/aspect_ratio,
    legend.position = "bottom",
    plot.background = element_rect(fill = "white", color = "white")
    )

ggsave("thumbnail.png", us_tz_map, width = 5 * aspect_ratio, height = 5)
us_tz_map

And what we see is there are a bunch of redundant times zone specification, especially in the Midwest.

US_tz %>%
  select(zone, latitude, longitude) %>%
  arrange(longitude) %>%
  gt() %>%
  opt_stylize(style = 6, color = "blue", add_row_striping = TRUE) %>%
  tab_header(title = "Countries with Multiple TZs")
Countries with Multiple TZs
zone latitude longitude
America/Adak 52.66667 -177.13333
America/Nome 64.56667 -165.78333
Pacific/Honolulu 21.71667 -158.35000
America/Anchorage 61.30000 -149.91667
America/Yakutat 60.35000 -140.35000
America/Sitka 57.75000 -135.41667
America/Juneau 58.41667 -134.60000
America/Metlakatla 55.73333 -132.15000
America/Los_Angeles 34.18333 -118.80000
America/Boise 44.41667 -116.35000
America/Phoenix 34.33333 -112.46667
America/Denver 40.08333 -105.03333
America/North_Dakota/Beulah 48.10000 -102.43333
America/North_Dakota/Center 48.08333 -102.23333
America/North_Dakota/New_Salem 47.53333 -102.05000
America/Menominee 45.56667 -88.45000
America/Indiana/Vincennes 39.30000 -88.23333
America/Indiana/Petersburg 39.00000 -87.98333
America/Chicago 41.85000 -87.65000
America/Indiana/Tell_City 38.13333 -87.43333
America/Indiana/Knox 42.03333 -87.11667
America/Indiana/Marengo 38.90000 -87.01667
America/Indiana/Winamac 41.13333 -86.78333
America/Indiana/Indianapolis 39.86667 -86.63333
America/Kentucky/Louisville 38.50000 -86.31667
America/Kentucky/Monticello 37.60000 -85.78333
America/Indiana/Vevay 39.60000 -85.10000
America/Detroit 43.20000 -83.78333
America/New_York 41.55000 -74.38333

Citation

BibTeX citation:
@online{sinks2023,
  author = {Sinks, Louise E.},
  title = {Tidy {Tuesday:} {Daylight} {Savings} {Time}},
  date = {2023-03-28},
  url = {https://lsinks.github.io/posts/2023-03-28-tidytuesday-timezones/tidytuesday-timezones.html},
  langid = {en}
}
For attribution, please cite this work as:
Sinks, Louise E. 2023. “Tidy Tuesday: Daylight Savings Time.” March 28, 2023. https://lsinks.github.io/posts/2023-03-28-tidytuesday-timezones/tidytuesday-timezones.html.