Repository files navigation

Extendingbroom to time series forecasting

Thesweep package extends thebroom tools (tidy, glance, andaugment) for performing forecasts and time series analysis in the“tidyverse”. The package is geared towards “tidying” the forecastworkflow used with Rob Hyndman’sforecast package.

• Designed for modeling and scaling forecasts using the thetidyverse tools inR for Data Science
• Extendsbroom for model analysis (ARIMA, ETS, BATS, etc)
• Tidies theforecast objects for easy plotting and “tidy” datamanipulation
• Integratestimetk to enable dates and datetimes (irregular timeseries) in the tidied forecast output

The package contains the following elements:

1. model tidiers:sw_tidy,sw_glance,sw_augment,sw_tidy_decomp functions extendtidy,glance, andaugmentfrom thebroom package specifically for models (ets(),Arima(),bats(), etc) used for forecasting.

2. forecast tidier:sw_sweep converts aforecast object to atibble that can be easily manipulated in the “tidyverse”.

sweep enables converting aforecast object totibble. The resultis ability to usedplyr,tidyr, andggplot natively to manipulate,analyze and visualize forecasts.

Often forecasts are required on grouped data to analyse trends insub-categories. The good news is scaling from one time series to many iseasy with the varioussw_ functions in combination withdplyr andpurrr.

A common goal in forecasting is to compare different forecast modelsagainst each other.sweep helps in this area as well.

If you are familiar withbroom, you know how useful it is forretrieving “tidy” format model components.sweep extends this benefitto theforecast package workflow with the following functions:

• sw_tidy: Returns model coefficients (single column)
• sw_glance: Returns accuracy statistics (single row)
• sw_augment: Returns residuals
• sw_tidy_decomp: Returns seasonal decompositions
• sw_sweep: Returns tidy forecast outputs.

The compatibility chart is listed below.

Function Compatibility

Here’s how to get started.

Development version with latest features:

# install.packages("remotes")remotes::install_github("business-science/sweep")

Thesweep package includes several vignettes to help users get up tospeed quickly:

• SW00 - Introduction tosweep
• SW01 - Forecasting Time Series Groups in the tidyverse
• SW02 - Forecasting Using Multiple Models