News

Development

New features:

• modelsummary() gets a gof_function argument which accepts functions to extract custom information from models.

datasummary_balance():

• formula can now include variables on the left-hand side to indicate the subset of columns to summarize: datasummary_balance(mpg + hp ~ am, data = mtcars) Thanks to @etiennebacher for feature request #751.

Bugs:

• escape argument not respected in datasummary_df(). Thanks to @adamaltmejd for report #740
• datasummary_correlation() supports data.table. Thanks to volatilehead on Twitter for report #737.
• Accepts named estimate argument when using shape and statistics are horizontal. Thanks to @iago-pssjd for report #745.
• Labelled values but no label for variable broke datasummary(). Thanks to @marklhc for report #752.
• coef_map does not work when there is a group. Thanks to @mccarthy-m-g for report #757.

2.0.0

MAJOR BREAKING CHANGE: The default output format is now tinytable instead of kableExtra. Learn more about tinytable here:

https://vincentarelbundock.github.io/tinytable/

To revert to the previous behavior persistently, users can call:

library(modelsummary) config_modelsummary(factory_default = “kableExtra”)

Other breaking changes:

• The statistic_override argument was replaced by vcov over 1 year ago, with appropriate deprecation warnings. It is now fully removed.
• The group argument was replaced by shape several releases ago. It is now fully removed.
• datasummary_skim()
  • histograms are only available with the tinytable backend. This allows a lot of code simplification and more customization.
  • The order of arguments type and output is switched for consistency with other functions.
  • histogram argument is deprecated.

New features:

• datasummary_skim():
  • New type="all" by default to display both numeric and categorical variables in a single table with distinct panels. This feature is only available with the tinytable backend (default).
  • by argument allows summarizing numeric variables by group.
  • fun_numeric argument accepts a list of functions to control the summary columns.
• modelsummary():
  • statistic and estimate can be specified as named vectors to control the names of statistics when displayed in different columns using the shape argument. (Thanks to @mps9506 for bug report #722)
  • modelsummary(panels, shape = "cbind") automatically adds column-spanning labels when panels is a named nested list of models.
• config_modelsummary() gets a startup_message argument to silence the startup message persistently.
• Improved documentation and vignettes, providing clearer instructions and examples.
• Updated tests and snapshots to ensure reliability and consistency across changes.

Bug fixes:

• Fixed Issue #399: datasummary_balance() siunitx formatting.
• Fixed Issue #782: Useless warning in some modelplot() calls. Thanks to @iago-pssjd for the report and @florence-laflamme for the fix.
• Addressed various bugs and made optimizations for better performance and user experience.

1.4.5

• tinytable supports histograms in datasummary_skim()
• config_modelsummary() supports tinytable factory.

1.4.4

• Support the tinytable package as an output format (“factory”): https://vincentarelbundock.github.io/tinytable/
• Quarto: md output format is recognized.
• options(modelsummary_factory_default) is respected, even in qmd->md documents.

Bugs:

• Some omitted coefficients with I() operator in formulas. Issue #693.

1.4.3

Misc:

• Duplicate values in shape groups are removed automatically for cleaner labels.
• “Title” line no longer indented in markdown tables. Thanks to Ryan Briggs for report #671.

Bugs:

• Small p values were not displayed properly in HTML output using kableExtra. Issue #669.

1.4.2

New:

• Minimal support for Typst output, with auto-detection in Quarto documents.
• strip argument in dvnames.
• s.value statistic is now available whenever p.value is available. See Greenland (2019).
• datasummary_skim() now includes histograms in gt tables.

Bugs:

• GOF term names get escaped in LaTeX. Thanks to @shreyasgm for reviving Issue #546.
• Conflict with furrr generated errors for some models. Thanks to @sammo3182 for Issue #647.

1.4.1

New:

• fmt_sci() can now be used in the fmt argument for rounding with scientific notation.

Bugs:

• Group separators respect add_rows with shape="rbind". Thanks to @lrose1 for Report #626.
• Bad column with horizontal models in shape and grouped estimates. Thanks to @iago-pssjd for Report #631.
• coef_rename=TRUE bug with grouped estimates. Thanks to @iago-pssjd for Report #631.
• Upstream issue #881 in parameters meant that vcov was no longer used for confidence intervals.

1.4.0

• Built-in support for markdown tables.
• Package no longer depends on kableExtra. Recommends an additional install for other formats.
• Persistent configuration of default output format: config_modelsummary(factory_default = "gt")
• shape = "rcollapse" and shape = "rbind"
• glance_custom() can drop GOF by assigning NA: https://stackoverflow.com/questions/75215355/assigning-different-coefficient-names-and-goodness-of-fit-statistics-to-differen
• When a statistic is not available, modelsummary prints an empty cell instead of returning an error.
• “\label{tab:something}” works in title even when escape=TRUE
• Multiple fixest_multi objects supported.
• options(modelsummary_future = FALSE) disables future parallelism.

Bug fixes:

• statistic=NULL is now respected when shape="rbind". Thanks to Panos Mavros for report #620.
• get_estimates() supports vcov string shortcuts and formulas. Thanks to @ethans-carl for report #605.
• Quarto and Rmarkdown documents include situnix in header automatically for decimal alignement with align="ddd"
• escape is now respected by modelsummary with shape="rbind". Thanks to @chickymonkeys for report #622.

1.3.0

Breaking change:

• The default column label style in modelsummary() has changed from “Model 1” to “(1)”. The benefits are: labels are no longer in English by default; use less horizontal space; eliminate the “Model” redundancy. Unfortunately, this could break code in some edge cases where users rely on column names to manipulate tables. The old behavior can be restored by calling: options(modelsummary_model_labels="model")

New features:

• shape="rbind" to stack multiple regression tables and create “panels” with labelled groups of models.
• fmt: new helper functions for different formatting styles
  • fmt = fmt_decimal(2): decimal digits
  • fmt = fmt_decimal(digits = 2, pdigits = 4): decimal digits with p value-specific setting
  • fmt = fmt_sprintf("%.3f"): sprintf() decimal
  • fmt = fmt_sprintf("%.3e"): sprintf() scientific
  • fmt = fmt_significant(3): significant digits
  • fmt = fmt_statistic("estimate" = 2, "std.error" = 3): statistic-specific formatting
  • fmt = fmt_term("(Intercept)" = 2, "hp" = 3): term-specific formatting
  • fmt = fmt_identity(): raw values
• New styles for default column labels in modelsummary, such as Roman Numerals or letters in parentheses.
  • Set the style with a global option: options(modelsummary_model_labels = "roman")
  • Supported styles: “model”, “arabic”, “letters”, “roman”, “(arabic)”, “(letters)”, “(roman)””
• modelplot(draw = FALSE) now returns a p.value column. This allows conditional aesthetics (see the modelplot vignette).
• Better integration with the marginaleffects package.

Bugs:

• Some fixest models returns useless “group.x” and “group.y” columns. Isse #591. Thanks to Adam Altmejd for the report.

1.2.0

Breaking change:

• With the shape and output="dataframe" arguments, there always used to be a group column. Now, this column has the same name as the variable in the shape formula (“response”, “component”, etc.).

New features:

• shape can include multiple groups.
• coef_rename can be an unnamed vector of length equal to the number of terms in the final table, obtained after coef_map and coef_omit are applied and models are merged.
• coef_omit accepts numeric indices. Positive values: coefficients to omit. Negative values: coefficients to keep.
• datasummary_skim: Increased maximum number of variables to 250.
• Quarto notebooks compile to Word and Markdown automatically.

Bug fixes:

• Order of notes preserved in some output format (Issue #577)

1.1.0

Breaking change:

• Requires siunitx version 3.0.25 LaTeX package.
• The title argument now respects the escape argument for all kableExtra output formats. This can break tables in which users manually escaped titles.

New features:

• “d” is accepted for decimal-alignment in the align argument for all output formats. modelsummary(mod, align = "ld")
• New update_modelsummary() function makes it easy to install the dev versions of modelsummary and its dependencies (mostly useful for Vincent and people who report bugs).
• Rounding: display at least one significant digit by default.
• Automatic renaming of haven labels in modelsummary(), datasummary(), datasummary_skim()
• Allow output = "filename.csv"
• Allow output = "filename.xlsx"
• add_columns argument supported in modelsummary()
• datasummary_balance supports the stars argument.
• Allow stars and confidence intervals with align = "d" column.

Bug fixes:

• In some locales, the HTML minus sign created problems in the output. We only use it in “known” locales.
• Many minor bug fixes

1.0.2

• Minor release to fix CRAN failure

1.0.1

• shape argument accepts interactions with the colon “:” character. This combines two columns into one, which can be useful to display terms and group names in a single column.
• Parallelization using parallel::mclapply. See ?modelsummary
• modelsummary no longer computes confidence intervals when not necessary, which can save some time. Also see: conf_level=NULL
• Added log likelihood to GOF for lm and glm models.
• Removed extraneous warnings
• Bug fixes

1.0.0

This first major release accompanies the publication of an article in the Journal of Statistical Software:

Arel-Bundock, Vincent (2022). “modelsummary: Data and Model Summaries in R.” Journal of Statistical Software, 103(1), 1-23. doi:10.18637/jss.v103.i01 https://doi.org/10.18637/jss.v103.i01.’

If you like modelsummary, please cite the JSS article and tell your friends about it.

Minor changes:

• gof_map="all" includes all available statistics. gof_map="none" excludes all statistics.
• Bug fixes

0.11.1

• Better printout for term names in mixed-effects models
• {brms} and {stanreg} models now extracted with diagnostic=NULL and test=NULL by default for speed.

0.11.0

Breaking changes:

• modelsummary_wide is no longer available. Use the shape argument of modelsummary instead.
• modelsummary now uses the easystats packages (performance and parameters) to extract estimates and goodness-of-fit statistics instead of broom. This can be reverted by setting a global option: options(modelsummary_get="broom"). This change aims to (1) increase consistency across models, (2) improve the developers’ ability to push bug fixes upstream when necessary, and (3) improve support for mixed effects, bayesian, and GAM models. The two main drawbacks are: (a) The set of printed statistics may be slightly different from previous versions of modelsummary (b) The group identifiers used in the shape formula will also be different for certain models (e.g., in nnet::multinom, y.level becomes response).

New features:

• The shape argument accepts a formula and can reshape information in myriad ways. Deprecates the group argument. Examples:
  • ~ statistic: statistics are shown horizontally in distinct columns.
  • model ~ term: models in rows and terms in columns.
  • term + y.level + statistic ~ model: grouped coefficients for multivariate outcome in nnet::multinom
  • y.level ~ model: partial match is the same as the previous formula
• Format distinct statistics differently by passing a named list to fmt:
  • modelsummary(mod, fmt = list(estimate = 2, std.error = 1, rmse = 4))
• Use glue to apply functions to numeric values by setting fmt = NULL. Example:
  • modelsummary(model, fmt = NULL, estimate = "{log(estimate)}")
• Update for breaking changes after fixest 0.10.4

Bug fixes:

• group_map rename issue
• Residual standard error mistakenly labelled “RMSE” in lm models.
• datasummary_skim output to jpg should now works
• escape fixes

0.10.0

• New exponentiate argument for modelsummary() and modelplot()
• gof_map accepts a vector such as c("rmse", "nobs", "r.squared")
• Drop rlang dependency
• Bug fixes

0.9.6

datasummary_balance:

• Accepts ~ 1 as a formula to summarize all data.

Misc:

• documentation improvements
• RMSE included by default in models of class lm

0.9.5

modelsummary:

• vcov strings like HC1 and Robust are now case-insensitive
• gof_map now accepts a data.frame or tibble with a fmt list-column which includes functions (see Examples in docs)
• R2 is no longer computed by default for bayesian and mixed effects models. An informative one-time warning is printed about the metrics argument.

datasummary_skim:

• Histograms now work in Jupyter
• Bugfix: harmless error message is no longer printed

kableExtra factory:

• The col.names argument can now be passed to kableExtra::kbl through the … ellipsis.

Misc:

• Many small improvements to the vignettes and docs
• output = "github_document" is now supported

0.9.4

• Bug fix: siunitx and rounding NA

0.9.3

modelsummary:

• F statistic takes into account vcov argument
• Support group = group ~ model + term

datasummary_balance:

• Weighted means and standard deviations are now supported. Counts and percentages are not, but raise a warning.

Misc:

• Bugfix: rounding in LaTeX w/ siunitx and NaN entries.
• output=‘jupyter’ no longer prints an extraneous TRUE to the notebook

0.9.2

modelsummary:

• Improved vcov argument handling for fixest models (#357 by @grantmcdermott)
• Fix display of fixest::i() variables and interactions (#361 by @grantmcdermott)
• Consistent display of clustered SEs (#356, #363 and #366 by @grantmcdermott)

datasummary_correlation:

• add_rows and add_columns arguments are now available here.

Misc:

• Global options for output factories are renamed: modelsummary_factory_default, modelsummary_factory_html, etc.
• Hot fix for change in R-devel behavior or intersect

Bug fixes:

• datasummary_balance: escape variable names when escape=TRUE
• Blogdown LaTeX dependency bug when output is HTML

0.9.1

Breaking change:

• Support for dcolumn for dot-aligned columns is deprecated. Use “d” in the align argument instead.

Other changes:

• LaTeX output: Numeric entries are wrapped in the \num{} function from the siunitx package by default. This produces much nicer formatting. This can be disabled with a global option. See ?modelsummary
• The align argument accepts a “d” column for dot-alignment using the siunitx LaTeX package: align="ldd".
• HTML tables display proper minus signs.
• New escape argument in most table-building functions.
• LaTeX output accepts the threeparttable=TRUE argument through ...
• No more dependency on tidyr

modelsummary:

• group: The order of terms in the formula determines the order of rows/columns

modelsummary_wide:

• Note: This function will eventually be deprecated
• Bugfix with statistic=NULL.

modelplot:

• Preserves order of models in the user-supplied list

datasummary_crosstab:

• statistic=NULL produces a very basic crosstab

datasummary_crosstab:

• Default alignment “lrrrrr” consistent with other datasummary_* functions

0.8.1

modelsummary:

• Disable stars footnote with options("modelsummary_stars_note" = FALSE)
• longtable=TRUE works for LaTeX output
• Interactions with “:” are no longer converted to “x” when coef_map or coef_rename are used.
• group = model ~ term + group is now supported.

datasummary_skim:

• datasummary_skim("categorical") keeps NA by default. Users can convert variables to factors before calling datasummary_skim to exclude NA.

Other:

• Improved warnings for bad calls: modelsummary(model1, model2)
• gt titles use the new caption argument in the gt 0.3.0 function
• Bug fix: Overaggressive tests for glue strings prevented functions inside {}

0.8.0

Breaking change:

• The default significance markers stars=TRUE have been updated to be consistent with the default output from base R (e.g., in summary.lm). The new significance thresholds are: “+” p < 0.1, “” p < 0.05, ”” p < 0.01, ”” p < 0.001

datasummary_crosstab:

• New function to produce cross-tabulations

datasummary:

• N is smart enough to return either the number of elements in a subset or the number of non-missing observations in a variable

datasummary_balance:

• Keeps NAs in factor variables by default. Users can convert their variables with the factor() function to omit NAs automatically.

modelsummary:

• themes can be set using global options (experimental)
• new vcov options: “bootstrap”, “HAC”, “NeweyWest”, “Andrews”, “panel-corrected”, “weave”, “outer-product”
• A valid get_gof (glance) is now optional.
• … is pushed through to sandwich, which allows things like: modelsummary(model, vcov = "bootstrap", R = 1000, cluster = "firm")

Other:

• Jupyter notebook support via output="jupyter"
• Bug fixes

0.7.0

modelsummary:

• new arguments for modelsummary: group and group_map for grouped parameters (e.g., outcome levels in multinomial logit or components of gamlss model).
• dvnames() makes it easy to get dependent variable column titles (thanks to @NickCH-K)
• output="modelsummary_list" to save a lightweight list-based representation of the table which can be saved and fed to modelsummary once more to get a full table.
• vcov adds a row to note the type of standard errors.
• modelsummary accepts a single model with multiple vcovs.
• get_gof forwards … to model_performance
• coef_map accepts unnamed vectors for easy subsetting
• fixest::fixest_multi support
• options(modelsummary_get) to set the order of extraction functions to use under the hood (broom vs. easystats vs. all)
• metrics argument of performance::model_performance is available via modelsummary’s … ellipsis to limit the GOF statistics in Bayesian models.
• users can omit the stars legend note by using glue strings: estimate="{estimate}{stars}"
• output=“html” can use gt by setting options(modelsummary_factory_html="gt")

datasummary_correlation:

• passes ... forward
• new function: datasummary_correlation_format
• datasummary_correlation’s method argument accepts functions and “pearspear” (thanks to @joachim-gassen)

datasummary:

• datasummary functions and rounding accept …, big.mark, etc.

datasummary_skim:

• now works with haven_labeled numeric
• faster tables with bayesian models.

Bug fixes and lints

0.6.6

• new output format: latex_tabular
• tidy_custom allows partial term matches
• modelsummary(coef_rename) accepts functions
• new function coef_rename for use in modelsummary(coef_rename=coef_rename)
• modelplot accepts add_rows to add reference categories
• informative error message when estimate or statistic is not available
• bug fixes

0.6.5

• statistic_override becomes vcov
• vcov accepts shortcuts: “robust”, “stata”, “HC0”, etc.
• vcov accepts formulas for clustered SEs: ~group
• modelsummary_wide has a new “stacking” argument
• html horizontal rule to separate estimates form gof
• gof_map accepts list of lists. only needs 3 columns.
• support officedown Rmd
• estimate accepts a vector for per model estimates
• options(modelsummary_default) can be markdown, html, latex
• bug: passing arguments through …
• bug: stars and rounding

0.6.4

• glue format for estimate and statistic
• easystats support for model info extraction
• deprecate statistic_vertical
• deprecate extract_models. Use modelsummary(output=“dataframe”) instead.
• modelplot pushes … through to modelsummary(output=“dataframe”)
• datasummary_skim(type=“dataset”)
• gof_map omits by default
• datasummary_balance uses row percentages
• statistic_override does not require a list
• statistic_override accepts a single model
• N function for well formatted N in datasummary
• Bug fixes

0.6.3

• new function: modelsummary_wide
• coef_omit and gof_omit use grepl(perl=TRUE)
• fmt accepts integer, string or function and respects options(OutDec=“,”)
• align argument for modelsummary
• align is more liberal to accept dcolumn alignment
• glance_custom methods for lfe and fixest
• bug fixes

0.6.2

• new argument: coef_rename
• new function: datasummary_df
• preserve term order in modelsummary
• refactor datasummary_balance
• datasummary_skim uses svg histograms instead of unicode
• removed 5 dependencies
• pass … to kableExtra::kbl for more customization
• test improvements
• internal code style
• bug fixes

0.6.1

• bug fixes

0.6.0

• default HTML output factory is now kableExtra
• interaction “:” gsubbed by “0d7”
• dependencies: removed 1 depends, 3 imports, and 3 suggests
• word_document knitr works out-of-the-box
• bug fixes

0.5.1

• glance_custom.fixest ships with modelsummary

0.5.0

• datasummary
• datasummary_skim
• datasummary_balance
• datasummary_correlation
• modelplot
• allow duplicate model names
• bug: can’t use coef_map with multiple statistics (thanks @sbw78)
• bug: wrong number of stars w/ statistic=‘p.value’ (thanks @torfason)
• output=‘data.frame’. extract is no longer documented.

0.4.1

• add_rows now accepts a data.frame with “position” and “section” columns
• add_rows_location is deprecated
• bug in sanity_output prevented overwriting files

0.4.0

• huxtable support
• flextable support
• estimate argument
• fixest tidiers
• website and vignette improvements
• gof_map additions
• glance_custom
• tidy_custom

0.3.0

• Out-of-the-box Rmarkdown compilation to HTML, PDF, RTF
• kableExtra output format for LaTeX and Markdown
• Support for threeparttable, colors, and many other LaTeX options
• Deprecated arguments: filename, subtitle
• Deprecated functions: clean_latex, knit_latex
• pkgdown website and doc improvements
• mitools tidiers
• New tests

0.2.1

• Convenience function to render markdown in row/column labels
• bug: breakage when all GOF were omitted
• Clean up manual with @keywords internal
• bug: tidyr import

0.2.0

• gt is now available on CRAN
• new latex_env argument for knit_latex and clean_latex
• bug when all gof omitted
• bug in statistic_override with functions
• bug caused by upstream changes in tab_style
• bug caused by upstream changes in filename=‘rtf’
• Allow multiple rows of uncertainty estimates per coefficient
• Preserve add_rows order
• Display uncertainty estimates next to the coefficient with statistic_vertical = FALSE
• Better clean_latex function
• Can display R2 and confidence intervals for mice-imputed lm-models
• Internal functions have @keywords internal to avoid inclusion in docs
• Statistic override accepts pre-formatted character vectors

0.1.0

• Initial release (gt still needs to be installed from github)