Introduction to prettyglm

When working with Generalized Linear Models it is often useful to create informative and beautiful summaries of the fitted model coefficients.This document introduces prettyglms’s main sets of functions, and shows you how to apply them.

Data: titanic

To explore the functionality of prettyglm we will use the titanic data set to perform logistic regression. This data set was sourced from https://www.kaggle.com/c/titanic/data and contains information about passengers aboard the titanic, and a target variable which indicates if they survived.

data('titanic')
head(titanic) %>%
  knitr::kable(table.attr = "style='width:30%;'" ) %>% 
  kableExtra::kable_classic(full_width = T, position = "center")

PassengerId	Survived	Pclass	Name	Sex	Age	SibSp	Ticket	Fare	Cabin	Embarked	Cabintype
1	0	3	Braund, Mr. Owen Harris	male	22	1	A/5 21171	7.2500	Missing	S	Missing
2	1	1	Cumings, Mrs. John Bradley (Florence Briggs Thayer)	female	38	1	PC 17599	71.2833	C85	C	C
3	1	3	Heikkinen, Miss. Laina	female	26	0	STON/O2. 3101282	7.9250	Missing	S	Missing
4	1	1	Futrelle, Mrs. Jacques Heath (Lily May Peel)	female	35	1	113803	53.1000	C123	S	C
5	0	3	Allen, Mr. William Henry	male	35	0	373450	8.0500	Missing	S	Missing
6	0	3	Moran, Mr. James	male	NA	0	330877	8.4583	Missing	Q	Missing

Pre-processing

A critical step for this package to work well is to set all categorical predictors as factors.

# Easiest way to convert multiple columns to a factor.
columns_to_factor <- c('Pclass',
                       'Sex',
                       'Cabin', 
                       'Embarked',
                       'Cabintype')
titanic  <- titanic  %>%
  dplyr::mutate_at(columns_to_factor, list(~factor(.)))

Building a glm

For this vignette we will use stats::glm() but prettyglm also supports parsnip and workflow model objects which use the glm model engine.

survival_model <- stats::glm(Survived ~ Pclass + Sex + Age + Fare + Embarked + SibSp + Parch + Cabintype, 
                             data = titanic, 
                             family = binomial(link = 'logit'))

Create pretty table of model coefficients with `pretty_coefficients()`

pretty_coefficients() allows you to create a pretty table of model coefficients, including categorical base levels.

pretty_coefficients(survival_model)

Variable	Level	Estimate	Std.error	P.Value
(Intercept)	(Intercept)	4.9902304	0.8148416	0
Pclass	1	0.0000000	0.0000000	NA
	2	-0.7001117	0.5071900	0.16747
	3	-1.8257623	0.5207283	0.00045
Sex	female	0.0000000	0.0000000	NA
Sex	male	-2.6858619	0.2280753	0
Age	Age	-0.0437710	0.0085211	0
Fare	Fare	0.0029054	0.0029893	0.33108
Embarked	C	0.0000000	0.0000000	NA
	Q	-0.7941971	0.6048757	0.18919
	S	-0.4336321	0.2836360	0.12631
SibSp	SibSp	-0.3546154	0.1305480	0.0066
Parch	Parch	-0.0698581	0.1250851	0.57651
Cabintype	A	0.0000000	0.0000000	NA
	B	-0.5672106	0.8047623	0.48092
	C	-1.1929795	0.7656414	0.1192
	D	-0.1838547	0.8120140	0.82088
	E	0.4033056	0.8153736	0.62086
	F	0.1140862	1.1176106	0.91869
	G	-1.8615347	1.3320706	0.16227
	Missing	-1.1033884	0.7893559	0.16216
	T	-13.5702506	535.4115815	0.97978
Goodness-of-Fit: AIC: 653.6 , Devience : 617.6 , Null Devience: 960.9

You can also complete a type III test on the coefficients by specifying a type_iii argument.

pretty_coefficients(survival_model, type_iii = 'Wald')

Variable	Level	Estimate	Std.error	P.Value	Type.III.P.Value
(Intercept)	(Intercept)	4.9902304	0.8148416	0	0
Pclass	1	0.0000000	0.0000000	NA	0
	2	-0.7001117	0.5071900	0.16747
	3	-1.8257623	0.5207283	0.00045
Sex	female	0.0000000	0.0000000	NA	0
Sex	male	-2.6858619	0.2280753	0	0
Age	Age	-0.0437710	0.0085211	0	0
Fare	Fare	0.0029054	0.0029893	0.33108	0.33108
Embarked	C	0.0000000	0.0000000	NA	0.22964
	Q	-0.7941971	0.6048757	0.18919
	S	-0.4336321	0.2836360	0.12631
SibSp	SibSp	-0.3546154	0.1305480	0.0066	0.0066
Parch	Parch	-0.0698581	0.1250851	0.57651	0.57651
Cabintype	A	0.0000000	0.0000000	NA	0.10512
	B	-0.5672106	0.8047623	0.48092
	C	-1.1929795	0.7656414	0.1192
	D	-0.1838547	0.8120140	0.82088
	E	0.4033056	0.8153736	0.62086
	F	0.1140862	1.1176106	0.91869
	G	-1.8615347	1.3320706	0.16227
	Missing	-1.1033884	0.7893559	0.16216
	T	-13.5702506	535.4115815	0.97978
Goodness-of-Fit: AIC: 653.6 , Devience : 617.6 , Null Devience: 960.9

or return a data frame by setting return_data to TRUE.

Create beautiful plots of variable relativities with `pretty_relativities()`

pretty_relativities() creates an interactive duel axis plot (via plotly), which plots the fitted relativity and the number of records in that category.

pretty_relativities(feature_to_plot= 'Embarked',
                    model_object = survival_model)

## Warning: `arrange_()` is deprecated as of dplyr 0.7.0.
## Please use `arrange()` instead.
## See vignette('programming') for more help
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_warnings()` to see where this warning was generated.

prettyglm-vignette