library(tidyverse)
data(hdv2003, package = "questionr")
d <-
  hdv2003 |> 
  mutate(
    sexe = sexe |> fct_relevel("Femme"),
    groupe_ages = age |>
      cut(
        c(18, 25, 45, 65, 99),
        right = FALSE,
        include.lowest = TRUE,
        labels = c("18-24 ans", "25-44 ans",
                   "45-64 ans", "65 ans et plus")
      ),
    etudes = nivetud |> 
      fct_recode(
        "Primaire" = "N'a jamais fait d'etudes",
        "Primaire" = "A arrete ses etudes, avant la derniere annee d'etudes primaires",
        "Primaire" = "Derniere annee d'etudes primaires",
        "Secondaire" = "1er cycle",
        "Secondaire" = "2eme cycle",
        "Technique / Professionnel" = "Enseignement technique ou professionnel court",
        "Technique / Professionnel" = "Enseignement technique ou professionnel long",
        "Supérieur" = "Enseignement superieur y compris technique superieur"
    ) |> 
    fct_na_value_to_level("Non documenté")  
  )

library(labelled)
d <- d |> 
  set_variable_labels(
    sport = "Pratique un sport ?",
    sexe = "Sexe",
    groupe_ages = "Groupe d'âges",
    etudes = "Niveau d'études",
    relig = "Rapport à la religion",
    heures.tv = "Heures de télévision / jour"
  )

library(gtsummary)
theme_gtsummary_language(
  "fr",
  decimal.mark = ",",
  big.mark = " "
)

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    )
  )
tbl_desc

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    statistic = all_categorical() ~ "{p}% ({n}/{N})"
  )
tbl_desc

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    statistic = all_categorical() ~ "{p}% ({n}/{N})",
    percent = "row"
  )
tbl_desc

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    statistic = all_categorical() ~ "{p}% ({n}/{N})",
    percent = "row",
    digits = all_categorical() ~ c(1, 0, 0)
  )
tbl_desc

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    statistic = all_categorical() ~ "{p}% ({n}/{N})",
    percent = "row",
    digits = all_categorical() ~ c(1, 0, 0)
  ) |> 
  modify_column_hide("stat_1")
tbl_desc

tbl_desc <-
  d |> 
  tbl_summary(
    by = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    statistic = all_categorical() ~ "{p}% ({n}/{N})",
    percent = "row",
    digits = all_categorical() ~ c(1, 0, 0)
  ) |> 
  modify_column_hide("stat_1") |> 
  modify_header("stat_2" ~ "**Pratique d'un sport**")
tbl_desc

tbl_uni <-
  d |>
  tbl_uvregression(
    y = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    method = glm,
    method.args = list(family = binomial)
  )
tbl_uni

tbl_uni <-
  d |>
  tbl_uvregression(
    y = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    method = glm,
    method.args = list(family = binomial),
    exponentiate = TRUE
  )
tbl_uni

tbl_uni <-
  d |>
  tbl_uvregression(
    y = sport,
    include = c(
      sexe, groupe_ages, etudes,
      relig, heures.tv
    ),
    method = glm,
    method.args = list(family = binomial),
    exponentiate = TRUE
  ) |> 
  modify_column_hide("stat_n")
tbl_uni

mod <- glm(
  sport ~ sexe + groupe_ages + etudes + relig + heures.tv,
  family = binomial,
  data = d
)
mod

mod <- glm(
  sport ~ sexe + groupe_ages + etudes + relig + heures.tv,
  family = binomial,
  data = d
)
mod2 <- step(mod, trace = 0)
mod2

mod <- glm(
  sport ~ sexe + groupe_ages + etudes + relig + heures.tv,
  family = binomial,
  data = d
)
mod2 <- step(mod, trace = 0)
tbl_multi <-
  mod2 |> 
  tbl_regression()
tbl_multi

mod <- glm(
  sport ~ sexe + groupe_ages + etudes + relig + heures.tv,
  family = binomial,
  data = d
)
mod2 <- step(mod, trace = 0)
tbl_multi <-
  mod2 |> 
  tbl_regression(exponentiate = TRUE)
tbl_multi

list(tbl_desc, tbl_uni, tbl_multi) |> 
  tbl_merge()

list(tbl_desc, tbl_uni, tbl_multi) |> 
  tbl_merge(
    tab_spanner = c(
      NA,
      "**Régressions univariées**",
      "**Régression multivariée**"
    )
  )

list(tbl_desc, tbl_uni, tbl_multi) |> 
  tbl_merge(
    tab_spanner = c(
      NA,
      "**Régressions univariées**",
      "**Régression multivariée**"
    )
  ) |> 
  bold_labels()