In earowang/tsibble: Tidy Temporal Data Frames and Tools

tsibble

knitr::opts_chunk$set(
  collapse = TRUE, comment = "#>", fig.path = "man/figures/"
)
options(tibble.print_min = 5)

The tsibble package provides a data infrastructure for tidy temporal data with wrangling tools. Adapting the tidy data principles, tsibble is a data- and model-oriented object. In tsibble:

1. Index is a variable with inherent ordering from past to present.
2. Key is a set of variables that define observational units over time.
3. Each observation should be uniquely identified by index and key.
4. Each observational unit should be measured at a common interval, if regularly spaced.

Installation

You could install the stable version on CRAN:

install.packages("tsibble")

You could install the development version from Github using

# install.packages("remotes")
remotes::install_github("tidyverts/tsibble")

Get started

Coerce to a tsibble with as_tsibble()

To coerce a data frame to tsibble, we need to declare key and index. For example, in the weather data from the package nycflights13, the time_hour containing the date-times should be declared as index, and the origin as key. Other columns can be considered as measured variables.

library(dplyr)
library(tsibble)
weather <- nycflights13::weather %>% 
  select(origin, time_hour, temp, humid, precip)
weather_tsbl <- as_tsibble(weather, key = origin, index = time_hour)
weather_tsbl

The key can be comprised of empty, one, or more variables. See package?tsibble and vignette("intro-tsibble") for details.

The interval is computed from index based on the representation, ranging from year to nanosecond, from numerics to ordered factors. The table below shows how tsibble interprets some common time formats.

| Interval | Class | |--------------|---------------------------| | Annual | integer/double | | Quarterly | yearquarter | | Monthly | yearmonth | | Weekly | yearweek | | Daily | Date/difftime | | Subdaily | POSIXt/difftime/hms |

A full list of index classes supported by tsibble can be found in package?tsibble.

fill_gaps() to turn implicit missing values into explicit missing values

Often there are implicit missing cases in time series. If the observations are made at regular time interval, we could turn these implicit missingness to be explicit simply using fill_gaps(), filling gaps in precipitation (precip) with 0 in the meanwhile. It is quite common to replaces NAs with its previous observation for each origin in time series analysis, which is easily done using fill() from tidyr.

full_weather <- weather_tsbl %>%
  fill_gaps(precip = 0) %>% 
  group_by_key() %>% 
  tidyr::fill(temp, humid, .direction = "down")
full_weather

fill_gaps() also handles filling in time gaps by values or functions, and respects time zones for date-times. Wanna a quick overview of implicit missing values? Check out vignette("implicit-na").

index_by() + summarise() to aggregate over calendar periods

index_by() is the counterpart of group_by() in temporal context, but it groups the index only. In conjunction with index_by(), summarise() aggregates interested variables over time periods. index_by() goes hand in hand with the index functions including as.Date(), yearweek(), yearmonth(), and yearquarter(), as well as other friends from lubridate. For example, it would be of interest in computing average temperature and total precipitation per month, by applying yearmonth() to the index variable (referred to as .).

full_weather %>%
  group_by_key() %>%
  index_by(year_month = ~ yearmonth(.)) %>% # monthly aggregates
  summarise(
    avg_temp = mean(temp, na.rm = TRUE),
    ttl_precip = sum(precip, na.rm = TRUE)
  )

While collapsing rows (like summarise()), group_by() and index_by() will take care of updating the key and index respectively. This index_by() + summarise() combo can help with regularising a tsibble of irregular time space too.

Learn more about tsibble

An ecosystem, the tidyverts, is built around the tsibble object for tidy time series analysis.

• The tsibbledata package curates a range of tsibble data examples to poke around the tsibble object.
• The feasts package provides support for visualising the data and extracting time series features.
• The fable package provides common forecasting methods for tsibble, such as ARIMA and ETS. The fabletools package, which is fable built upon, lays the modelling infrastructure to ease the programming with tsibble.

---

Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.

earowang/tsibble documentation built on Feb. 6, 2024, 11:27 a.m.