Lubridate

Clever name aside, lubridate is a fantastic way (arguably the BEST way) to work with date and datetime objects. Lubridate allows you to easily convert strings or numbers to date-times, deal with timezones, and perform math with date-times.

Example

Start by loading the lubridate and tidyverse libraries along with your data.

# load the tidyverse and lubridate libraries
library(tidyverse)
## -- Attaching packages -------------------------------------------- tidyverse 1.3.0 --
## v ggplot2 3.2.1     v purrr   0.3.3
## v tibble  2.1.3     v dplyr   0.8.3
## v tidyr   1.0.0     v stringr 1.4.0
## v readr   1.3.1     v forcats 0.4.0
## -- Conflicts ----------------------------------------------- tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()
library(lubridate)
## 
## Attaching package: 'lubridate'
## The following object is masked from 'package:base':
## 
##     date
# load in the csv for analysis
csv <- "time_series_covid_19_confirmed.csv"
coronavirus_data <- data.frame(read.csv(csv, header = TRUE, stringsAsFactors = FALSE))

Clean your data so that it follows a tidy format.

# clean the dataset; pivot to long form, separate data column,
# create confirmed column (based on grouping), and eliminate duplicates
coronavirus_data <- coronavirus_data %>%
  pivot_longer(c("X1.22.20":"X3.27.20"),
    names_to = "Date",
    values_to = "Cases") %>%
  separate("Date", into = c("X", "Date"), sep = "X") %>%
  group_by(Country.Region, Date) %>%
  mutate("Confirmed" = sum(Cases))%>%
  distinct(Country.Region, Date, Confirmed)

The first great feature of lubridate is it’s ability to convert strings to dates. This can be done using any combination of ymd.

# convert Date column to a date object
coronavirus_data$Date <- mdy(coronavirus_data$Date)

You can also create time intervals by using the interval function. This creates a bounded timeline with options to check for overlapping, if one interval is contained within the other, and much more!

# find US time interval
us_cases <- coronavirus_data %>%
  filter(Country.Region == "US") %>%
  filter(Confirmed > 0)

us_start <- us_cases$Date[1]
us_end <- us_cases$Date[length(us_cases$Date)]
us_interval <- interval(us_start, us_end)

Create two more intervals.

# find Italy time interval
italy_cases <- coronavirus_data %>%
  filter(Country.Region == "Italy") %>%
  filter(Confirmed > 0)

italy_start <- italy_cases$Date[1]
italy_end <- italy_cases$Date[length(italy_cases$Date)]
italy_interval <- interval(italy_start, italy_end)

# find Spain time interval
spain_cases <- coronavirus_data %>%
  filter(Country.Region == "Spain") %>%
  filter(Confirmed > 0)

spain_start <- spain_cases$Date[1]
spain_end <- spain_cases$Date[length(spain_cases$Date)]
spain_interval <- interval(spain_start, spain_end)

You can check if any of the intervals overlap.

# check interval overlaps
int_overlaps(us_interval,italy_interval)
## [1] TRUE
int_overlaps(us_interval,spain_interval)
## [1] TRUE
int_overlaps(spain_interval,italy_interval)
## [1] TRUE

You can also examine the length of an interval and the difference in their lengths.

# examine the length of an interval
as.duration(us_interval)
## [1] "5616000s (~9.29 weeks)"
as.period(us_interval)
## [1] "2m 5d 0H 0M 0S"
# what is the difference in length of the outbreak in the US vs Spain?
dseconds(setdiff(us_interval,spain_interval))
## [1] "864000s (~1.43 weeks)"

Finally, you can see if an interval is contained entirely within a different interval.

# check if an interval is contained within another
spain_interval %within% us_interval
## [1] TRUE
us_interval %within% spain_interval
## [1] FALSE

Lastly, there is a bit of math you can do with lubridate. For example, you can add exactly one month to any given date. Here, I found the last day of January, February, and March for 2020 (notice that lubridate takes into account that 2020 is a leap year).

# find the last day of each month using lubridate math
jan_end <- ymd("2019-12-31") %m+% months(1)
feb_end <- ymd("2019-12-31") %m+% months(2)
march_end <- ymd("2019-12-31") %m+% months(3)

# get total cases by month
us_by_month <- us_cases %>%
  mutate("Month" = month(Date, label = TRUE)) %>%
  group_by(Month) %>%
  filter(Date == jan_end | Date == feb_end | Date == march_end | Date == us_end)

Conclusion

Lubridate is an extremely powerful resource for dealing with dates and times. Beyond what was shown here, the user is able to deal with times, timezones, daylight savings time, and much more.