- add correlation to
explore_cor()
when usinggeom_points
- add parameter
nthread
toexplain_xgboost()
. (#45) - add
interact()
. (#47) - add interactivity as default in explore shiny app. (#47)
- add
create_data_abtest()
. - add basic color parameter to
explore()
&abtest()
functions. - add
get_color()
- move NA-info in
explore()
from title to subtitle. (#48) - add more descriptive infos into
explore()
subtitle. - add
color
parameter forexplore()
,explore_*()
,report()
- add
bins
parameter totarget_explore_num()
mix_color()
with one color as parameter generates colors from light to darktarget_explore_num()
bar positioning changes from max to mean value
- fix undefined / not meaningful values in abtest() shiny-app
- rename vignettes
explore_*.Rmd
toexplore-*.Rmd
- add
explain_xgboost()
(#42) - add
drop_var_by_names()
(#43) - add
drop_var_not_numeric()
(#43) - add
drop_var_low_variance()
(#43) - add
drop_var_no_variance()
(#43) - add
drop_var_with_na()
(#43) - add
drop_obs_with_na()
(#43) - add
drop_obs_if()
(#43) - add
mix_color()
- add
show_color()
- add
create_data_esoteric()
create_data_empty()
has no longer a parameterseed
- add
check_vec_low_variance()
(internal helper function)
- Fix CRAN check warning in running example
get_nrow()
(#41)
- explore no longer imports MASS and RandomForest. for
explain_logreg()
andexplain_forest()
, you will receive a prompt to install these packages in interactive sessions. (#2 1, @olivroy)
- add
explain_forest()
. - add
predict_target()
. - add
create_data_newsletter()
. - add
use_data_beer()
anduse_data_starwars()
functions (#20, #23) - explore now has a website. (#17, #19, @olivroy)
abtest()
now supports numeric target (t-test).abtest_targetpct()
with count data (parametern
).abtest()
andexplore()
can now run without data (shiny app). If no data are provided,palmerpenguins::penguins
is used. (#25)- New vignettes were added. (#27, #28, #29, #31)
- Documentation and examples were enhanced. (#38, @olivroy, #32, #33, #36, #37)
create_data_()
use_data_*()
return data sets as tibble.
- drop use of
fct_explicit_na()
(forcats >= 1.0.0) and uselinewidth
for ggplot2 (>= 3.4.0) (deprecated) (#15, @olivroy) - explore no longer depends on assertthat, tidyr, and broom. (#16, #21, @olivroy).
- Error messages are now more informative (#18, @olivroy)
- explore now uses testthat 3rd edition and GitHub actions (#16, @olivroy)
- explore has been re-licensed as MIT (#22)
add_var_random_01()
creates variable of type integer- add
target_name
&factorise_target
parameter to morecreate_data_*()
- add
target1_prob
parameter to morecreate_data_*()
- add checks to
create_data_*()
- format variable random_moon
- add
abtest()
- remove native pipe in
explore_tbl()
- fix error in
explore()
median ifNA
values - add tests for
explore()
(no error if data containsNA
)
- Switch back to
%>%
in vignettes (compatibility R < 4.1) (#6) - No hard coded path for markdown-templates
- Add
create_data_unfair()
create_data_app()
gains ascreen_size
argument.
- Dependency DT (>= 0.3.0)
- Improve and use native pipe in README.
- Improve documentation
- Add function
create_data_app()
- Add support for integer64
- Bugfix
report()
>100 variables - No warning-message in
explore_count()
- Redesign
explore_tbl()
- Add mean to
explore_density()
plot - Add
create_data_churn()
- Add
add_var_random_moon()
- Vignettes: switch from
%>%
to|>
- Add
create_notebook_explore()
- rename create_x_data() to
create_data_x()
- rename add_x_var() to
add_var_x()
- extend
create_data_*()
functions - extend
add_var_*()
functions explain_tree()
: set defaultminsplit = 20
explain_tree()
: set prior probabilitiesexplore()
andreport()
:targetpct
as alternative tosplit
parameterbalance_target()
: add parameter seed- fix variable type for
create_data_x()
- using md in roxygen
- all
dwh_*()
functions are no longer included in {explore} Alternative: source https://github.com/rolkra/dwh - add
create_fake_data()
- add
create_random_data()
- add
add_random_var()
- add
get_var_buckets()
total_fig_height()
: parametersvar_name_target
,var_name_n
- code styling
- report templates: add var buckets (to plot large number of variables)
- change
theme_light()
intoindividual theme()
so thatset_theme
works. - add rmarkdown to Suggests in DESCRIPTION
- fix URL in DESCRIPTION
- drop LazyData in DESCRIPTION
- format DESCRIPTION
explain_tree()
gains aweights
parameter.- flip no/yes label in rpart.plot
- fix
minsplit
for count-data - add
weight_target()
- add
plot_legend_targetpct()
- add legend in targetpct rmarkdown-template
- add unit testing (testthat)
- Bugfix
explore_bar()
:NA
in plot explore_count()
: convert target into factorexplore_count()
: add default title (cat name)explore_count()
: add parameter numeric, max_cat, max_target_catexplain_tree()
: convert character variables into factors (count data)explain_tree()
: parameter out ("plot" | "model")explain_logreg()
: parameter out ("tibble" | "model")vignette("explore_titanic")
: change to tibblevignette("explore_mtcars")
: add explanations- change theme_minimal to theme_light
- dwh_fastload(): add parameters overwrite and append
- update README.md
- Fix Github URL
- new Vignette
vignette("explore_penguins")
- new Vignette
vignette("explore_titanic")
(count data) explore_count()
: plot count() output- add default parameter
n
for count data:explore()
,explore_all()
,explore_tbl()
,explain_tree()
,report()
,describe()
,describe_cat()
,describe_num()
,describe_tbl()
,total_fig_height()
explore_tree()
: default value for minsplit = 10% of obsexplore_cor()
: usegeom_point()
for small datasetsexplore_shiny()
: usebrowseURL()
with parameterbrowser=NULL
describe_tbl()
: add observations containingNA
guess_cat_num()
: parameter description (optional)count_pct()
: no renaming of variables.
Maintenance update:
- fix breaking changes tibble 3.0.0
Maintenance update:
- fix param
...
in description (PR#16223, see https://bugs.r-project.org/show_bug.cgi?id=16223)
explore_bar()
: add parameter numericdescribe_all()
returns a tibbledescribe_all()
: column 'variable' is character (not factor)report()
split = TRUE as default- add
rescale01()
- add parameter
rescale01
toclean_var()
- add function
count_pct()
- add
out='tibble'
todescribe_cat()
- add function
explore_targetpct()
- split source-code file into multiple files
format_num_auto()
without brackets- treat Date variables as cat
report()
fix automatic file extension .html- add
simplify_text()
- add parameter
simplify_text
toclean_var()
- fix link in README.md
Prepare for new dplyr 0.8.4 (#2, @romainfrancois)
- prepare
explore_tbl()
for dplyr 0.8.4 describe_num()
with default digits=6describe_cat()
bugfix variable with all NAdescribe_all()
bugfix variable with all NAexplain_tree()
bugfix dataframe with 0 rows- improve speed
describe()
text output (RMarkdown) explore()
now checks if data is a data.frame
Interactive data exploration now accept categorical and numerical targets (next to a binary target).
explain_tree()
: target can be bin/num/catexplain_tree()
: add parameter max_target_catexplore_shiny()
: target can be bin/num/cat- add function
format_num_auto()
total_fig_height()
replaces the now deprecatedget_nrow()
.- add parameter title to
explore_cor()
- add support for POSIXct in
describe()
- improved handling of dataframes with no observations
- add parameter
title
toexplore_density()
- add parameter
nvar
tototal_fig_height()
- update README.md
- update Vignettes
- add NEWS.md
- add hex sticker
Many functions now accept categorical and numerical targets (next to a binary target). If you want to force which geom is used for visualisation, you can use explore_bar() and explore_density()
. New function explore_tbl()
to visualise a dataframe/table (type of variables, number of NA, ...)
- add function
explore_bar()
explore_density()
now using correct tidy eval, target cat > 2 possibletarget_explore_cat()
now using correct tidy evaltarget_explore_num()
now using correct tidy evaladd plot_var_info()
- plots a info-text to a variable as ggplot obj.plot_var_info()
used in explore/explore_all ifplot_var_info()
used if explore empty data- add parameter
max_cat
inexplore_bar()
,explore_density()
andexplain_tree()
- add
explore_tbl()
- drop
explore_cat()
&explore_num()
- rename template_report_target_den.html > _split.html
- intelligent placing of labels in plots
- add info window "generating report ..." in
explore_shiny()
format_num()
-> format_num_kMB(), format_num_space()format_target()
-> if numeric split 0/1 by meanreport()
-> default .html file extension- consistency showing NA info in explore-title
- parameter split: default = FALSE
- allow numeric (num) target in explore_all & report
describe_tbl()
-> fix target if not bindescribe()
: change out="vector" to out="list"
- fix parameter in
explore()
:auto_scale
,na
- fix number of
NA
inexplore()
(move code beforeauto_scale
) explore_density()
with target: drop plot title "propensity by"explore_shiny()
: use output_dir /tempdir()
- change naming "attribute" to "variable" (consistent)