In andreas50/utsMultivariate: Multivariate Unevenly Spaced Time Series

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/"
)
options(width=100)
library(utsMultivariate)

Introduction

This package provides S3 classes ("uts_vector" and "uts_matrix") for multivariate unevenly spaced time series. The individual time series are uts objects.

R users familiar with vectors and matrices should find it straightforward to work with the corresponding time series objects, because the usage of many functions (such as [, [[, length, etc.) is completely analogous.

Installation

This package is not yet available on CRAN, but can be installled from GitHub, either using the R package devtools:

devtools::install_github("andreas50/uts", build_vignettes=TRUE)
devtools::install_github("andreas50/utsMultivariate")

or using the R package remotes:

remotes::install_github("andreas50/uts")
remotes::install_github("andreas50/utsMultivariate")

Sample Code

First, a few examples how to use a time series vector.

# Get sample unevenly-spaced time series vector
x <- ex_uts_vector()
x

# Plot the time series vector
par(mai=c(0.5, 0.5, 0.2, 0.2))
plot(x, type="o", cex.axis=0.8)

# Get first and last observation value(!) of each time series
first(x)
last(x)

# Get first and last observation time(!) of each time series
start(x)
end(x)

# Insert new observation
x[as.POSIXct("2007-11-10 10:00:00"), "apples"] <- 45

# Sample the time series at a specific time point, using one of several supported interpolation methods
sample_values(x, as.POSIXct("2007-11-10"), interpolation="linear")

# Get maximum observation value of each time series
max(x)

# Time series arithmetic
(x*2 + 5)$apples$values

# Apply a function to the individual time series of a time series vector
sapply(x, length)
sapply(x, range)

# Convert the time series vector to data.frame, i.e. flatten the data
as.data.frame(x)

# Conversely, convert data in flat (long) format to a time series vector
values <- data.frame(apples=1:10, oranges=letters[1:10], stringsAsFactors=FALSE)
uts_vector_wide(values, times=as.POSIXct("2015-01-01") + ddays(1:10))

# Convert data in wide format to a time series vector
values <- c(5, 6, 6, 7)
times <- as.POSIXct("2016-01-01") + dhours(1:4)
names <- c("a", "b", "a", "a")
uts_vector_long(values, times, names)

Most operatios for time series vectors also apply to a time series matrix, so only a few examples are given here.

# Create a time series matrix by recycling the elements of a time series vector
y <- uts_matrix(x, nrow=2, ncol=3)
rownames(y) <- c("foo", "bar")
colnames(y) <- c("red", "green", "blue")
y

# Extract first two rows, and second column (coming soon)
#y[, 1:2, "green"]

# Transpose
t(y)