Use Roxygen 7.2.1.
Fix errors on CRAN.
Updated renv and packages.
Updated GitHub Action for pkgdown.
predict_gam() now has the extra argument type, which allows the user to return the predicted values based on all terms or based on each term separately.Use renv for development.
@return roxygen entry added to all functions documentation.
ci_z argument in plot_difference().
get_smooths_difference() now includes a group column with a numeric index of significance blocks, so that plotting difference smooths with more than two alternating significance values is correct (see plot-smooths vignette, last example).✨ - get_smooths_difference() which returns a tibble with the difference of two levels of a smooth (#11).
🍇 - plot-smooths.Rmd now includes a plotting example with get_smooths_difference().
👷 - Add GitHub actions for R CMD check.
🖼 - Add logo!
reintroduced plot_difference().
imported code from itsadug for better integration with plot_difference() and compatibility with R 3.6.1.
🍇 include example of get_gam_predictions() in plot smooths vignette.
get_gam_predictions() now returns an .idx column which assigns a number to each curve (when multiple variables are used). The .idx column can be used to correctly group geom_ribbon() for plotting CIs.plot_difference() due to the archiving of itsadug. To plot difference smooths, you can use mgcViz::plotDiff().prevent error from dplyr::lag() when type of default is different (prepare for upcoming dplyr 1.0.0)
roxygen version 7.1.0
transform argument in get_gam_predictions() and plot_smooths() (closes https://github.com/stefanocoretta/tidymv/issues/9)
exlude_random argument in plot_difference() (it was TRUE by default and it was not possible to change it)
error message in get_gam_predictions when using discretised bam models and exclude_random = TRUE (which is the default)
two example datasets to be used in the examples
example of non-Gaussian GAM in plot*smooths vignette
support for models with non*syntactic column names (log(y) ~ s(log(x)))
add option of setting values to NULL in predict_gam() when excluding terms to reduce computation time (also mentioned in the vignette)
improved performance of get_gam_predictions() when excluding terms
included mention to loaded packages in the vignettes
wrong examples in plot_smooths() and plot_difference() (closes https://github.com/stefanocoretta/tidymv/issues/10)
handling of s(bs = "re") smooths
alpha of 0 line in difference smooth plot (now set to 0.5)
prepared for CRAN first release
plot_difference()time_series is now deprecated and replaced with series. time_series will be removed in future releases.predict_gam() to return a dataframe with all predictors and fitted values with standard error
geom_smooth_ci() which provides a new ggplot2 geom to conveniently plot smooths and confidence intervals from the output of predict_gam()
vignette that illustrates how to use the new functions
examples in the documentation
create_event_start() (> create_start_event(), it breaks backward compatibility)plot_gamsd() (use plot_smooths() and plot_difference())install_github argument for building vignettesrlangseries_length default to 100 in plot_difference()plot_difference() to plot difference smooths
support for plotting single smooth
plot_gamsd() now mentions plot_difference()split argument for separating columns (useful for interactions)
how to plot interactions in vignette
exclude_terms argument to exclude terms when predicting
exclude tensor.smooths when plotting smooths
experimental ylim argument to plot_gamsd()
plot_gamsd() which said "will be deprecated"get_gam_predictions()get_gam_predictions() for predicting with a gam object
plot_smooths() for plotting smooths
plot*smooths.Rmd vignette
support for models with smooths with terms not in time_series
typo in documentation of create_event_start()
double call of itsadug in DESCRIPTION
note about lag on check
note about non*imported itsadug
incompatibility with [email protected]
plot_gamsd() will be deprecated: use plot_smooths instead. Plotting both smooths and difference smooth is not supported yet.rm_re parameter for removing random effects in plot_gamsd
depends on itsadug
event.col to event_col in create_start_event() for style consistencyplot_gamsd()gam model in the vignette now has a reference smooth s(x2)
updated vignette
lag() for create_event_start()plot_gamsd about missing xmin and xmax in annotate if there is no differencecreate_start_event function
plot_gamsd function