working-with-dates-and-times.Rmd

---
title: "Working with dates and times"
description: |
  Introduction to lubridate for dates and times in R.
author:
  - name: Stephanie Hicks
    url: https://stephaniehicks.com/
    affiliation: Department of Biostatistics, Johns Hopkins
    affiliation_url: https://www.jhsph.edu
date: 10-05-2021
output:
  distill::distill_article:
    self_contained: false
    toc: true
    toc_depth: 3
categories:
  - module 2
  - week 6
  - tidyverse
  - R 
  - programming
  - dates and times
---


```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, error=TRUE, message=FALSE, warning=FALSE)
```

# Pre-lecture materials

### Read ahead

:::resources

**Before class, you can prepare by reading the following materials:**

1. https://r4ds.had.co.nz/dates-and-times.html

:::


### Acknowledgements

Material for this lecture was borrowed and adopted from

- https://jhu-advdatasci.github.io/2018/lectures/09-dates-times.html
- https://r4ds.had.co.nz/dates-and-times.html

# Learning objectives

:::keyideas

**At the end of this lesson you will:**

- Recognize the `Date`, `POSIXct` and `POSIXlt` class types in R to represent dates and times
- Learn how to create date and time objects in R using functions from the `lubridate` package
- Learn how dealing with time zones can be frustrating `r emojifont::emoji("scream_cat")` but hopefully less so after today's lecture `r emojifont::emoji("smiley_cat")` 
- Learn how to perform arithmetic operations on dates and times 
- Learn how plotting systems in R "know" about dates and times to appropriately handle axis labels 

:::


# Introduction

In this lesson, we will learn how to work with dates and times in R. 
These may seem simple as you use them all of the time in your day-to-day life, 
but the more you work with them, the more complicated they seem to get. 

Dates and times are hard because they have to reconcile two physical phenomena (the rotation of the Earth and its orbit around the sun) with a whole raft of geopolitical phenomena including months, time zones, and daylight savings time (DST). 

This lesson will not teach you every last detail about dates and times, but it will give you a solid grounding of practical skills that will help you with common data analysis challenges.

R has developed a special representation of dates and times

- Dates are represented by the `Date` class

- Times are represented by the `POSIXct` or the `POSIXlt` class

- Dates are stored internally as the number of days since 1970-01-01

- Times are stored internally as the number of seconds since
  1970-01-01

### The `lubridate` package

Here, we will focus on the `lubridate` R package, which makes it easier to work with dates and times in R. 

**check out the cheat sheet** at https://lubridate.tidyverse.org

A few things to note about it: 

- It largely replaces the default date/time functions in base R

- It contains methods for date/time arithmetic

- It handles time zones, leap year, leap seconds, etc.

`lubridate` is installed when you install `tidyverse`, but it is not loaded when you load `tidyverse`.
Alternatively, you can install it separately. 

```{r, eval=FALSE}
install.packages("lubridate") 
```

```{r}
library(tidyverse)
library(lubridate) 
```


# Creating date/times

There are three types of date/time data that refer to an instant in time:

- A **date**. Tibbles print this as `<date>`.

- A **time** within a day. Tibbles print this as `<time>`.

- A **date-time** is a date plus a time: it uniquely identifies an instant in time (typically to the nearest second). Tibbles print this as `<dttm>`. Elsewhere in R these are called `POSIXct`. 

In this lesson, we will focus on dates and date-times as R does not have a native class for storing times. If you need one, you can use the [`hms` package](https://cran.r-project.org/web/packages/hms/index.html).

You should always use the simplest possible data type that works for your needs. That means if you can use a date instead of a date-time, you should. Date-times are substantially more complicated because of the need to handle time zones, which we’ll come back to at the end of the lesson.

To get the current date or date-time you can use `today()` or `now()` from `lubridate`:

```{r}
today()

now()
```

Otherwise, there are three ways you are likely to create a date/time:

- From a string.
- From individual date-time components.
- From an existing date/time object.

They work as follows.

### From a string

Dates are of the `Date` class. 

```{r}
x <- today()
class(x)
```

Dates can be coerced from a character strings using some helper functions from `lubridate`. 
They automatically work out the format once you specify the order of the component. 

To use the helper functions, identify the order in which year, month, and day appear in your dates, then arrange "y", "m", and "d" in the same order. 

That gives you the name of the `lubridate` function that will parse your date. For example:

```{r}
ymd("1970-01-01")
ymd("2017-01-31")
mdy("January 31st, 2017")
dmy("31-Jan-2017")
```


:::keyideas 

**Pro tips**: 

- When reading in data with `read_csv()`, you may need to read in as character first and then convert to date/time

- `Date` objects have their own special `print()` methods that will always format as "YYYY-MM-DD"

- These functions also take unquoted numbers. 

```{r}
ymd(20170131)
```

:::

#### Alternate Formulations

Different locales have different ways of formatting dates

```{r}
ymd("2016-09-13")  ## International standard
ymd("2016/09/13")  ## Just figure it out
mdy("09-13-2016")  ## Mostly U.S.
dmy("13-09-2016")  ## Europe
```

All of the above are valid and lead to the exact same object.

Even if the individual dates are formatted differently, `ymd()` can usually figure it out.

```{r}
x <- c("2016-04-05", 
       "2016/05/06",
       "2016,10,4")
ymd(x)
```

### From individual date-time components

Sometimes the date components will come across multiple columns in a dataset. 

```{r}
library(nycflights13)

flights %>% 
  select(year, month, day)
```

To create a date/time from this sort of input, use `make_date()` for dates, or `make_datetime()` for date-times:

```{r}
flights %>% 
  select(year, month, day) %>% 
  mutate(departure = make_date(year, month, day))
```

:::questions

The `flights` also contains a `hour` and `minute` column. Use `make_datetime()` to create a date-time column called `departure`:

```{r}
# try it yourself

```

:::

### From other types

You may want to switch between a date-time and a date. That is the job of `as_datetime()` and `as_date()`:

```{r}
today()
as_datetime(today())

now()
as_date(now())
```

# Date-Times in R

### From a string

`ymd()` and friends create dates. To create a `date-time` from a character string, add an underscore and one or more of "h", "m", and "s" to the name of the parsing function:

Times can be coerced from a character string with `ymd_hms()`

```{r}
ymd_hms("2017-01-31 20:11:59")
mdy_hm("01/31/2017 08:01")
```

You can also force the creation of a date-time from a date by supplying a timezone:

```{r}
ymd_hms("2016-09-13 14:00:00")
ymd_hms("2016-09-13 14:00:00", tz = "America/New_York")
ymd_hms("2016-09-13 14:00:00", tz = "")

```

### `POSIXct` or the `POSIXlt` class

Let's get into some hairy details about date-times. Date-times are represented using the `POSIXct` or the `POSIXlt` class in R. What are these things? 

#### `POSIXct`

`POSIXct` is just a very large integer under the hood. It is a useful class when you want to store times in something like a data frame. 

Technically, the `POSIXct` class represents the number of **seconds** since 1 January 1970. (In case you were wondering, "POSIXct" stands for "Portable Operating System Interface", calendar time.)

```{r}
x <- ymd_hm("1970-01-01 01:00")
class(x) 
unclass(x)

typeof(x)

attributes(x)
```


#### `POSIXlt`

`POSIXlt` is a `list` underneath and it stores a bunch of other useful information like the day of the week, day of the year, month, day of the month

```{r}
y <- as.POSIXlt(x)

typeof(y)

attributes(y)
```

:::keyideas 

**Pro tips**: `POSIXlt`s are **rare** inside the tidyverse. They do crop up in base R, because they are needed to extract specific components of a date, like the year or month. 

Since `lubridate` provides helpers for you to do this instead, you do not really need them imho. 

`POSIXct`'s are always easier to work with, so if you find you have a `POSIXlt`, you should always convert it to a regular data time `lubridate::as_datetime()`.

:::

# Time Zones!

Time zones were created to make your data analyses more difficult as a data analyst.
Here are a few fun things to think about: 

* `ymd_hms()` function will by default use Coordinated Universal Time (UTC) as the time zone. UTC is the primary time standard by which the world regulates clocks and time.

You can go to Wikipedia to find the [list of time zones](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones)


* Specifying `tz = ""` in one of the `ymd()` and friends functions will use the local time zone

```{r}
x <- ymd_hm("1970-01-01 01:00", tz = "")
x

attributes(x)
```

:::keyideas 

**Pro tip**: The `tzone` attribute is optional. It controls how the time is printed, not what absolute time it refers to.

```{r}
attr(x, "tzone") <- "US/Pacific"
x

attr(x, "tzone") <- "US/Eastern"
x
```

:::


A few other fun things to think about related to time zones: 

* Almost always better to specify time zone when possible to avoid ambiguity

* Daylight savings time (DST)

* Some states are in two time zones

* Southern hemisphere is opposite


# Operations on Dates and Times

### Arithmetic 

You can add and subtract dates and times. You can do comparisons too (i.e. `==`, `<=`)

```{r, error=TRUE}
x <- ymd("2012-01-01", tz = "")  ## Midnight
y <- dmy_hms("9 Jan 2011 11:34:21", tz = "")
x - y ## this works
x < y ## this works
x > y ## this works
x == y ## this works
x + y ## what??? why does this not work? 
```

:::keyideas

**Remember**: `POSIXct` objects are a measure of seconds from an origin, usually the UNIX epoch (1st Jan 1970). 

Just add the requisite number of seconds to the object:

```{r}
x + 3*60*60 # add 3 hours
x
```

:::

Same goes for days. For example, you can just keep the date portion using `date()`: 

```{r}
y <- date(y)
y
```

And then add a number to the date (in this case 1 day)

```{r}
y + 1  
```

Cool eh? 


### Leaps and Bounds

Even keeps track of leap years, leap seconds, daylight savings, and time zones.

Leap years

```{r}
x <- ymd("2012-03-01")
y <- ymd("2012-02-28")
x - y
```

Not a leap year 

```{r}
x <- ymd("2013-03-01")
y <- ymd("2013-02-28")
x - y
```

BUT beware of time zones!

```{r}
x <- ymd_hms("2012-10-25 01:00:00", tz = "")
y <- ymd_hms("2012-10-25 05:00:00", tz = "GMT")
y - x
```

There are also things called [**leap seconds**](https://en.wikipedia.org/wiki/Leap_second).

```{r}
.leap.seconds
```

# Extracting Elements of Dates/Times

There are a set of helper functions in `lubridate` that can extract sub-elements of dates/times

### Date Elements

```{r}
x <- ymd_hms(c("2012-10-25 01:13:46",
               "2015-04-23 15:11:23"), tz = "")
year(x)
month(x)
day(x)
weekdays(x)
```

### Time Elements

```{r}
x <- ymd_hms(c("2012-10-25 01:13:46",
               "2015-04-23 15:11:23"), tz = "")
minute(x)
second(x)
hour(x)
week(x)
```



# Visualizing Dates


### Reading in the Data

```{r, warning=FALSE, message=FALSE}
library(here)
library(readr)
storm <- read_csv(here("data", "storms_2004.csv.gz"), progress = FALSE)
storm
```

```{r}
names(storm)
```


:::questions 

Let's take a look at the `BEGIN_DATE_TIME`, `EVENT_TYPE`, and `DEATHS_DIRECT` variables. 
Try to convert the `BEGIN_DATE_TIME` date/time column to a date/time R object.

```{r,message=FALSE}
library(dplyr)
storm %>% 
  select(BEGIN_DATE_TIME, EVENT_TYPE, DEATHS_DIRECT) %>% 
  mutate(date = dmy_hms(BEGIN_DATE_TIME))

# try it yourself

```

:::

Next, we do some wrangling to create a `storm_sub` data frame (code chunk set to `echo=FALSE` for the purposes of the lecture, but code is in the R Markdown).

```{r, echo=FALSE}
storm_sub <- 
  storm %>% 
  select(BEGIN_DATE_TIME, EVENT_TYPE, DEATHS_DIRECT) %>%
  mutate(begin = dmy_hms(BEGIN_DATE_TIME)) %>%
  rename(type = EVENT_TYPE, deaths = DEATHS_DIRECT) %>%
  select(begin, type, deaths)
```

```{r}
storm_sub
```


### Histograms of Dates/Times


We can make a histogram of the dates/times to get a sense of when storm events occur.

```{r}
library(ggplot2)
storm_sub %>%
  ggplot(aes(x = begin)) + 
  geom_histogram(bins = 20) + 
  theme_bw()
```

We can group by event type too.

```{r,fig.width=12,fig.height=12}
library(ggplot2)
storm_sub %>%
  ggplot(aes(x = begin)) + 
  facet_wrap(~ type) + 
  geom_histogram(bins = 20) + 
  theme_bw() + 
  theme(axis.text.x.bottom = element_text(angle = 90))
```

### Scatterplots of Dates/Times

```{r}
storm_sub %>%
  ggplot(aes(x = begin, y = deaths)) + 
  geom_point()
```

If we focus on a single month, the x-axis adapts.

```{r}
storm_sub %>%
  filter(month(begin) == 6) %>%
  ggplot(aes(begin, deaths)) + 
  geom_point()
```

Similarly, we can focus on a single day.

```{r}
storm_sub %>%
  filter(month(begin) == 6, day(begin) == 16) %>%
  ggplot(aes(begin, deaths)) + 
  geom_point()

```

# Summary

- Dates and times have special classes in R that allow for numerical and statistical calculations

- Dates use the `Date` class

- Date-Times (and Times) use the `POSIXct` and `POSIXlt` class

- Character strings can be coerced to Date/Time classes using the `ymd()` and `ymd_hms()` functions and friends. 

- The `lubridate` package is essential for manipulating date/time data

- Both `plot` and `ggplot` "know" about dates and times and will handle axis labels appropriately.



# Post-lecture materials

### Final Questions 

Here are some post-lecture questions to help you think about the material discussed.

:::questions

**Questions:**

1. What happens if you parse a string that contains invalid dates?

```{r, eval=FALSE}
ymd(c("2010-10-10", "bananas"))
```

2. What does the `tzone` argument to `today()` do? Why is it important?

```{r}
unclass(today())
```


3. Use the appropriate `lubridate` function to parse each of the following dates:

```{r}
d1 <- "January 1, 2010"
d2 <- "2015-Mar-07"
d3 <- "06-Jun-2017"
d4 <- c("August 19 (2015)", "July 1 (2015)")
d5 <- "12/30/14" # Dec 30, 20
```

4. Using the `flights` dataset, how does the distribution of flight times within a day change over the course of the year?

5. Compare `dep_time`, `sched_dep_time` and `dep_delay`. Are they consistent? Explain your findings.

:::


### Additional Resources 

:::resources 

- https://lubridate.tidyverse.org
- `lubridate` cheat sheet: https://rawgit.com/rstudio/cheatsheets/master/lubridate.pdf
- https://jhu-advdatasci.github.io/2018/lectures/09-dates-times.html
- https://r4ds.had.co.nz/dates-and-times.html

:::