R Create a Dummy Variable with the Day of Week

Question

I want to create 7 Dummy-Variables for all days of the week to my dataframe. My dataframe "BTC.USD" contains time series data of Bitcoin expressed in USD (Close = Closing Price):

# A tibble: 6 x 3
Date       Close Wday 
<date>     <dbl> <ord>
1 2015-12-31  430. Do   
2 2016-01-01  434. Fr   
3 2016-01-02  434. Sa   
4 2016-01-03  431. So   
5 2016-01-04  433. Mo   
6 2016-01-05  431. Di

(The abbreviations for Wday are in German)
How I can add 7 new columns with Dummy Variables for the days of the week? Is it better to use a map- or a apply-function?

You can do `df %>% spread(Wday, Wday) %>% mutate_at(3:8, funs(ifelse(is.na(.), 0, 1)))`. — tmfmnk, Jan 29 '19 at 08:36

RLave · Accepted Answer · 2019-01-29T08:47:17.730

You can use weekdays() to find the day of the week, then dcast().

tt <- "date close
2015-12-31  430   
2016-01-01  434   
2016-01-02  434   
2016-01-03  431   
2016-01-04  433   
2016-01-05  431"

d <- read.table(text=tt, header=T, stringsAsFactors = F)

library(dplyr) # for mutate() and `%>%`
library(data.table) # for dcast()
d %>% 
  mutate(date = as.Date(date)) %>% 
  mutate(weekday = weekdays(date)) %>% 
  dcast(date + close ~ weekday, fun.aggregate = length)

R Create a Dummy Variable with the Day of Week

1 Answers1