POL 478H1 F

class: center, middle, inverse, title-slide

# POL 478H1 F
## Dates and Time
### Olga Chyzh [www.olgachyzh.com]

---

## Package `lubridate`

- Read, format, and recode time variables

- Apply mathematical and other functions to time variables: e.g. add/subtract

- Convert time, e.g. weeks to months

---

## Variables in `Date` Format

Update the `classdata` package and load the `covid_daily` dataset.

```r
library(tidyverse)
library(magrittr)
library(devtools)
install_github("ochyzh/classdata")
library(classdata)
data("covid_daily")
str(covid_daily)
```

```
## tibble [526,323 x 9] (S3: tbl_df/tbl/data.frame)
##  $ fips         : num [1:526323] 1001 1001 1001 1001 1001 ...
##  $ county       : chr [1:526323] "Autauga County" "Autauga County" "Autauga County" "Autauga County" ...
##  $ date         : chr [1:526323] "2020-04-07" "2020-04-08" "2020-04-09" "2020-04-10" ...
##  $ state_name   : chr [1:526323] "Alabama" "Alabama" "Alabama" "Alabama" ...
##  $ population   : num [1:526323] 55200 55200 55200 55200 55200 55200 55200 55200 55200 55200 ...
##  $ confirmed    : num [1:526323] 12 12 15 17 19 19 19 23 24 26 ...
##  $ new_confirmed: num [1:526323] 0 0 3 2 2 0 0 4 1 2 ...
##  $ deaths       : num [1:526323] 1 1 1 1 1 1 1 1 1 1 ...
##  $ new_deaths   : num [1:526323] 1 0 0 0 0 0 0 0 0 0 ...
```

---

## Format Converter Functions

- Convert character-variables to date/time using `lubridate`'s converter functions:

- date: `ymd`, `mdy`, `dmy`, ...
    - time: `hm`, `hms`, ...
    - date and time: `ymd_hms`, `mdy_hm`, ...
- Order of letters specifies how the strings are parsed.

- Separators determined automatically.

---
## Examples

.pull-left[

```r
library(lubridate)
x<-"16-11-2020"
class(x)
```

```
## [1] "character"
```

```r
dmy(x)
```

```
## [1] "2020-11-16"
```

```r
class(dmy(x))
```

```
## [1] "Date"
```
]

.pull-right[

```r
y<-"Last updated at 10:30 am"
class(y)
```

```
## [1] "character"
```

```r
hm(y)
```

```
## [1] "10H 30M 0S"
```

```r
class(hm(y))
```

```
## [1] "Period"
## attr(,"package")
## [1] "lubridate"
```
]
---

## Or Apply to Convert Character Strings in a Dataset

```r
class(covid_daily$date)
```

```
## [1] "character"
```

```r
covid_daily$date1<-ymd(covid_daily$date)
class(covid_daily$date1)
```

```
## [1] "Date"
```

---
## Managing Date Variables

- Sumarise:

```r
summary(covid_daily$date)
```

```
##    Length     Class      Mode 
##    526323 character character
```

```r
summary(covid_daily$date1)
```

```
##         Min.      1st Qu.       Median         Mean      3rd Qu.         Max. 
## "2020-02-29" "2020-06-18" "2020-08-14" "2020-08-08" "2020-10-02" "2020-11-14"
```

- Extract month, year, ... using `month`, `year`, `week`, `wday`, `yday`, `hour`, `minute`

```r
covid_daily$month<-month(covid_daily$date1)
covid_daily$wday<-wday(covid_daily$date1, label=T)
```
---
## Durations

```r
mdy("10-16-2020")-mdy("10-02-2020")
```

```
## Time difference of 14 days
```

```r
mdy("10-16-2020")-months(1)
```

```
## [1] "2020-09-16"
```

```r
mdy("10-16-2020")-weeks(3)
```

```
## [1] "2020-09-25"
```

```r
mdy("10-16-2020")-days(13)
```

```
## [1] "2020-10-03"
```

---
## Your Turn

Use the daily Covid-19 data from the `classdata` package to plot the total cases in the US by month. Which months saw the highest spikes of cases?

---
## 7 Day Rolling Average

```r
library(zoo)
#Set theme options:
theme_set(theme_grey() + theme(panel.background = element_rect(fill = NA, color = 'black'))+ theme(axis.text=element_text(size=10),
					axis.title=element_text(size=12,face="bold")))
p<-covid_daily %>% group_by(date1) %>% summarise(sum_cases=sum(new_confirmed, na.rm=T), .groups="keep") %>% ungroup() %>% arrange(date1) %>% mutate(mean_cases7=zoo::rollmean(sum_cases,k=7,fill = NA)/1000) %>% ggplot(aes(x=date1,y=mean_cases7))+geom_line()+scale_y_continuous("Covid-19 Cases, Thousands")+scale_x_date("")

p1<-covid_daily %>% group_by(date1) %>% summarise(sum_deaths=sum(new_deaths, na.rm=T), .groups="keep") %>% ungroup() %>% arrange(date1) %>% mutate(mean_deaths7=zoo::rollmean(sum_deaths,k=7,fill = NA)) %>% ggplot(aes(x=date1,y=mean_deaths7))+geom_line()+scale_y_continuous("Covid-19 Deaths")+scale_x_date("")

library(gridExtra)
grid.arrange(p, p1, 
             ncol = 1, nrow = 2)
```
---
## 7 Day Rolling Average
<img src="09_dates_files/figure-html/unnamed-chunk-11-1.png" style="display: block; margin: auto;" />
---
## Days of the Week, Cases

```r
covid_daily  %>%  mutate(wday=wday(date1, label=T)) %>% group_by(wday) %>% ggplot(aes(x=wday,weight=new_confirmed/1000))+geom_bar()+scale_y_continuous("Covid-19 Cases")
```

<img src="09_dates_files/figure-html/unnamed-chunk-12-1.png" width="50%" style="display: block; margin: auto;" />
---
## Your Turn

Make a visualization to find out the day of the week with the most Covid-19 deaths.

---

## Use `sugrrants` to Visualize Temporal Data

```r
library(sugrrants)
calendar_df<-covid_daily %>% group_by(date1) %>% summarise(sum_cases=sum(new_confirmed, na.rm=T)/1000, .groups="keep") %>% frame_calendar(x=1, y=1, date=date1,calendar="monthly", week_start=7)

p<-calendar_df %>% ggplot(aes(x=.x,y=.y, group=date1)) + geom_tile(aes(fill = sum_cases), colour = "grey50") + geom_text(label=day(calendar_df$date1),size=3, nudge_x = .001, nudge_y=.02)+
  scale_fill_continuous("Number of Cases, thousands",low="white", high="red")
prettify(p, label=c("label"),label.padding = unit(0.2, "lines"))+
    theme(legend.position="bottom")+ggtitle("Covid-19 Cases in the US, 2020")+theme(plot.title = element_text(hjust = 0.5))
```
---
## Use `sugrrants` to Visualize Temporal Data

---
## Your Turn

Make a calendar visualization of Covid-19 deaths.