tsbox 0.3.1: extended functionality

Christoph Sax

The tsbox package provides a set of tools that are agnostic towards existing time series classes. The tools also allow you to handle time series as plain data frames, thus making it easy to deal with time series in a dplyr or data.table workflow.

Photo by James Sutton

Version 0.3.1 is now on CRAN and provides several bugfixes and extensions (see here for the full change log). A detailed overview of the package functionality is given in the documentation page (or in an older blog-post).

New and extended functionality

  • ts_frequency(): changes the frequency of a time series. It is now possible to aggregate any time series to years, quarters, months, weeks, days, hours, minutes or seconds. For low- to high-frequency conversion, the tempdisagg package can now convert low frequency to high frequency and has support for ts-boxable objects. E.g.:

      x <- ts_tbl(EuStockMarkets)
      #> # A tibble: 7,440 × 3
      #>   id    time                 value
      #>   <chr> <dttm>               <dbl>
      #> 1 DAX   1991-07-01 03:18:27  1629.
      #> 2 DAX   1991-07-02 13:01:32  1614.
      #> 3 DAX   1991-07-03 22:44:38  1607.
      #> 4 DAX   1991-07-05 08:27:43  1621.
      #> 5 DAX   1991-07-06 18:10:48  1618.
      #> # … with 7,435 more rows
      ts_frequency(x, "week")
      #> # A tibble: 1,492 × 3
      #>   id    time        value
      #>   <chr> <date>      <dbl>
      #> 1 DAX   1991-06-30  1618.
      #> 2 DAX   1991-07-07  1633.
      #> 3 DAX   1991-07-14  1632.
      #> 4 DAX   1991-07-21  1620.
      #> 5 DAX   1991-07-28  1616.
      #> # … with 1,487 more rows
  • ts_index(): returns an indexed series, with a value of 1 at the base period. This base period can now be specified more flexibly. E.g., the average of a year can defined as 1 (which is a common use case).

  • ts_na_interpolation(): A new function that wraps imputeTS::na_interpolation() from the imputeTS package and allows the imputation of missing values for any time series object.

  • ts_first_of_period(): A new function that replaces the date or time value by the first of the period. This is useful because tsbox usually relies on timestamps being the first of a period. The following monthly series has an offset of 14 days. ts_first_of_period() changes the timestamp to the first date of each month:

      x <- ts_lag(ts_tbl(mdeaths), "14 days")
      #> # A tibble: 72 × 2
      #>   time       value
      #>   <date>     <dbl>
      #> 1 1974-01-15  2134
      #> 2 1974-02-15  1863
      #> 3 1974-03-15  1877
      #> 4 1974-04-15  1877
      #> 5 1974-05-15  1492
      #> # … with 67 more rows
      #> # A tibble: 72 × 2
      #>   time       value
      #>   <date>     <dbl>
      #> 1 1974-01-01  2134
      #> 2 1974-02-01  1863
      #> 3 1974-03-01  1877
      #> 4 1974-04-01  1877
      #> 5 1974-05-01  1492
      #> # … with 67 more rows

Convert everything to everything

tsbox is built around a set of converters, which convert time series stored as ts, xts, data.frame, data.table, tibble, zoo, tsibble, tibbletime, tis, irts or timeSeries to each other:

x.ts <- ts_c(fdeaths, mdeaths)
x.xts <- ts_xts(x.ts)
x.df <- ts_df(x.xts)
x.dt <- ts_dt(x.df)
x.tbl <- ts_tbl(x.dt)
x.zoo <- ts_zoo(x.tbl)
x.tsibble <- ts_tsibble(x.zoo)
x.tibbletime <- ts_tibbletime(x.tsibble)
x.timeSeries <- ts_timeSeries(x.tibbletime)
x.irts <- ts_irts(x.tibbletime)
x.tis <- ts_tis(x.irts)
all.equal(ts_ts(x.tis), x.ts)
#> [1] TRUE

Use same functions for time series classes

Because this works reliably, it is easy to define a toolkit that works for all classes. So, whether we want to smooth, scale, differentiate, chain, forecast, regularize, impute or seasonally adjust a time series, we can use the same commands to whatever time series class at hand:

ts_trend(x.ts)   # estimate a trend line
ts_pc(x.xts)     # calculate percentage change rates (period on period)
ts_pcy(x.df)     # calculate percentage change rates (year on year)
ts_lag(x.dt)     # lagged series

There are many more. Because they all start with ts_, you can use auto-complete to see what’s around. Most conveniently, there is a time series plot function that works for all classes and frequencies:

  `Airline Passengers` = AirPassengers,
  `Lynx trappings` = ts_tis(lynx),
  `Deaths from Lung Diseases` = ts_xts(fdeaths),
  title = "Airlines, trappings, and deaths",
  subtitle = "Monthly passengers, annual trappings, monthly deaths"

time series plot

More posts

EFS vs. NFS for RStudio on Kubernetes (AWS): Configuration and considerations

Patrick Schratz

Accessing Google's API via OAuth2

Patrick Schratz

seasonal 1.9: Accessing composite output

Christoph Sax

Google Season of Docs with R: useR! Information Board

Ben Ubah

Running old versions of TeXlive with tinytex

Kirill Müller

tsbox 0.3.1: extended functionality

Christoph Sax

Celebrating one-year anniversary as RStudio Full Service Certified Partner

Cosima Meyer, Patrick Schratz

Deprecating a pkgdown site served via GitHub Pages

Patrick Schratz, Kirill Müller

gfortran support for R on macOS

Patrick Schratz

Seasonal Adjustment of Multiple Series

Christoph Sax

Dynamic build matrix in GitHub Actions

Kirill Müller

Setting up a load-balanced Jitsi Meet instance

Patrick Schratz

DevOps Expert (f/m/d, 60-100%)

cynkra team

Maintaining multiple identities with Git

Kirill Müller

Relational data models in R

Angel D'az, Kirill Müller

tempdisagg: converting quarterly time series to daily

Christoph Sax

tsbox 0.2: supporting additional time series classes

Christoph Sax

DevOps System Engineer (40-60%)

cynkra team

Introducing dm: easy juggling of tables and relations

Balthasar Sager

tsbox 0.1: class-agnostic time series

Christoph Sax

Time series of the world, unite!

Christoph Sax

Done “Establishing DBI”!?

Kirill Müller

Other blogs