Common Code 14 — Dates with lubridate

Parsing, extracting, arithmetic, and plotting time series

How to work with dates

Working with dates

Dates in R are their own data type — not text, not numbers. The lubridate package makes parsing, extracting, and doing arithmetic with dates simple and readable. This matters any time you work with field data, phenology, water quality records, or eBird data.

⬇️ Download the companion R script: 14_dates.R

library(tidyverse)
library(lubridate)
library(nycflights13)

---

1 · Parsing dates — turning strings into dates

The function name tells R the order of the components in your string:

ymd("2026-06-25")        # ISO format — international standard
mdy("06/25/2026")        # US format — common in Excel
dmy("25-06-2026")        # European format
ymd("20260625")          # no separator — also works

WarningThe most common date mistake

Dates imported from CSV or Excel often look like dates but are stored as text. glimpse() or str() will show them as <chr> not <date>. Always convert with ymd() / mdy() / dmy() after import — or use col_types = cols(date = col_date()) inside read_csv().

# Fix a date column that imported as character
my_data |>
  mutate(date = ymd(date_string))    # or mdy(), dmy() depending on format

Dates from Excel serial numbers

Excel stores dates as integers counting from 1900. as_date() handles this:

tibble(excel_date = c(45000, 45100, 45200)) |>
  mutate(real_date = as_date(excel_date, origin = "1899-12-30"))

---

2 · Extracting components

Once you have a proper <date> column, pull out whatever piece you need:

d <- ymd("2026-06-25")

year(d)              # 2026
month(d)             # 6
day(d)               # 25
yday(d)              # 176  — day of year (critical for phenology)
week(d)              # ISO week number
wday(d, label=TRUE)  # "Thu"
quarter(d)           # 2

TipDay of year (yday) is essential for ecology

Convert calendar dates to day of year before plotting seasonal patterns — it aligns across years and handles the year boundary cleanly. Ice-off dates, bird arrival dates, and bloom dates are all typically analysed as day of year.

Applied to a data frame

flights |>
  mutate(
    date       = make_date(year, month, day),
    month_name = month(date, label = TRUE, abbr = FALSE),
    day_of_yr  = yday(date),
    weekday    = wday(date, label = TRUE)
  ) |>
  select(flight, date, month_name, day_of_yr, weekday) |>
  head()

---

3 · Building dates from components

make_date(2026, 6, 25)                   # one date from three numbers
make_datetime(2026, 6, 25, 14, 30, 0)   # with hour, minute, second

# Build a date column from separate year/month/day columns
flights |>
  mutate(date = make_date(year, month, day)) |>
  select(flight, year, month, day, date) |>
  head()

---

4 · Date arithmetic

start <- ymd("2026-03-01")
end   <- ymd("2026-06-25")

end - start                    # 116 days (difftime)
as.numeric(end - start)        # 116 (plain number)

start + days(30)               # add 30 days
start + months(3)              # add 3 months (respects calendar)
start + years(1)               # add 1 year

Days between field visits

field_visits <- tibble(
  visit = 1:4,
  date  = ymd(c("2026-04-10","2026-05-08","2026-06-03","2026-07-01"))
) |>
  mutate(days_since_last = as.numeric(date - lag(date)))

field_visits

Notedays() vs ddays()

days(30) adds 30 calendar days. ddays(30) adds 30 × 86 400 seconds (a fixed duration). They differ at daylight-saving transitions. For ecological field data, use days().

---

5 · Filtering by date

flights |>
  mutate(date = make_date(year, month, day)) |>
  filter(date >= ymd("2013-07-01"),
         date <= ymd("2013-07-31"))   # July flights only

# Filter by extracted component
flights |>
  mutate(date = make_date(year, month, day),
         mon  = month(date)) |>
  filter(mon %in% c(12, 1, 2))        # winter months

---

6 · Plotting time series

Daily counts with x-axis formatted as dates

flights |>
  mutate(date = make_date(year, month, day)) |>
  count(date) |>
  ggplot(aes(x = date, y = n)) +
  geom_line(color = "steelblue", alpha = 0.8) +
  geom_smooth(method = "loess", span = 0.1,
              color = "tomato", se = FALSE) +
  scale_x_date(date_breaks = "1 month", date_labels = "%b") +
  labs(x = NULL, y = "Flights per day",
       title = "Daily flights from NYC — 2013") +
  theme_minimal()

scale_x_date() formats the x-axis for date objects. Common date_labels format codes:

Code	Output
%Y	2026
%b	Jun
%B	June
%m	06
%d	25
%b %Y	Jun 2026

---

7 · Ecological example — ice-off phenology

Is the lake freezing later or thawing earlier over time? Day of year is the right response variable:

ice_off <- tibble(
  year     = 1981:1995,
  date_str = c("1981-04-15","1982-04-08", ...)
) |>
  mutate(
    date      = ymd(date_str),
    day_of_yr = yday(date)
  )

ggplot(ice_off, aes(x = year, y = day_of_yr)) +
  geom_point(size = 3, color = "steelblue") +
  geom_smooth(method = "lm", se = TRUE,
              color = "tomato", fill = "tomato", alpha = 0.15) +
  labs(x = "Year", y = "Ice-off day of year",
       title = "Lake ice-off phenology 1981–1995") +
  theme_minimal()

---

Quick reference

Task	Code
Parse ISO date	ymd("2026-06-25")
Parse US date	mdy("06/25/2026")
Parse Euro date	dmy("25-06-2026")
Parse Excel serial	as_date(n, origin = "1899-12-30")
Build from parts	make_date(year, month, day)
Year / month / day	year(d) / month(d) / day(d)
Day of year	yday(d)
Day name	wday(d, label = TRUE)
Month name	month(d, label = TRUE, abbr = FALSE)
Add days	d + days(n)
Add months	d + months(n)
Days between	as.numeric(date2 - date1)
Format x-axis	scale_x_date(date_breaks="1 month", date_labels="%b")

---

End of Common Code 14 — Dates with lubridate. Next: Common Code 15 — Strings and factors.