5.1 Add Metadata for later filtering

Firstly we have to load a dataset into a dataframe:

# load data set
df <- read.csv("https://github.com/hslu-ige-laes/edar/raw/master/sampleData/centralOutsideTemp.csv",
               stringsAsFactors=FALSE,
               sep =";")

5.1.1 Year, Month, Day, Day of Week

To e.g. group, filter and aggregate data we need eventually the date splitted up in day, month and year:

library(dplyr)
library(lubridate)

 %>% %>% %>% %>% df$year <- as.Date(cut(df$time, breaks = "year"))
df$month <- as.Date(cut(df$time, breaks = "month"))
df$day <- as.Date(cut(df$time, breaks = "day"))
df$weekday <- wday(df$time,
                   label = TRUE,
                   locale = "English",
                   abbr = TRUE,
                   week_start = getOption("lubridate.week.start", 1))

This code first parses the timestamp with a specific timezone. Then three columns are added.

Please note that the month also contains the year and a day. This is useful for a later step where you can group the series afterwards.

head(df,2)

##                  time centralOutsideTemp
## 1 2018-03-21 11:00:00                5.2
## 2 2018-03-21 12:00:00                6.7

tail(df,2)

##                      time centralOutsideTemp
## 21864 2020-09-17 10:00:00              26.65
## 21865 2020-09-17 11:00:00              28.10

5.1.2 Hour, Minute, Second

df$hour <- as.POSIXct(lubridate::floor_date(df$time,"hours"), tz = "Europe/Zurich")
df$minute <- as.POSIXct(lubridate::floor_date(df$time,"minutes"), tz = "Europe/Zurich")
df$second <- as.POSIXct(lubridate::floor_date(df$time,"seconds"), tz = "Europe/Zurich")

5.1.3 Season of Year

For some analyses it is useful to color single points of a scatterplot according to the season. For this we need to have the season in a separate column:

library(redutils)
# get season from a date
getSeason(as.Date("2019-04-01"))

## [1] "Spring"

If you want to change the language, you can give the function dedicated names for the season:

getSeason(as.Date("2019-04-01"),
                  seasonlab = c("Winter","Frühling","Sommer","Herbst"))

## [1] "Frühling"

To apply this function to a whole dataframe we can use the dplyr mutate function. The code below creates a new column named “season”:

df <- dplyr::mutate(df, season = getSeason(df$time))

head(df,1)

##                  time centralOutsideTemp season
## 1 2018-03-21 11:00:00                5.2 Spring

tail(df,1)

##                      time centralOutsideTemp season
## 21865 2020-09-17 11:00:00               28.1 Summer