# ============================================================
# ANÁLISIS: PERCEPCIÓN SOBRE TRABAJO DE LA MUJER
# Base: Woman Work Perc.xlsx
# Estudiante: Nina María Sánchez Ramírez
# Tutor: Carlos Eduardo Alonso-Malaver
# Actividades:
# 1. Tablas univariadas
# 2. Tablas bivariadas
# 3. Análisis de Correspondencias Múltiples MCA
# ============================================================
# ============================================================
# 1. INSTALACIÓN Y CARGA DE PAQUETES
# ============================================================
paquetes <- c(
"readxl", "dplyr", "janitor", "FactoMineR",
"factoextra", "ggplot2", "openxlsx"
)
paquetes_no_instalados <- paquetes[!(paquetes %in% installed.packages()[, "Package"])]
if(length(paquetes_no_instalados) > 0){
install.packages(paquetes_no_instalados)
}
library(readxl)
library(dplyr)
##
## Adjuntando el paquete: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
library(janitor)
##
## Adjuntando el paquete: 'janitor'
## The following objects are masked from 'package:stats':
##
## chisq.test, fisher.test
library(FactoMineR)
library(factoextra)
## Cargando paquete requerido: ggplot2
## Welcome to factoextra!
## Want to learn more? See two factoextra-related books at https://www.datanovia.com/en/product/practical-guide-to-principal-component-methods-in-r/
library(ggplot2)
library(openxlsx)
# ============================================================
# 2. IMPORTAR BASE DE DATOS
# ============================================================
ruta <- "C:\\Users\\Nina Sanchez\\Documents\\ACER -LENOVO-NINA 2019-20\\1. USCO\\ESTADISTICA\\2026_Estadistica\\Malaver\\Woman Work Perc.xlsx"
datos <- read_excel(ruta)
# Revisar estructura inicial
str(datos)
## tibble [3,418 × 5] (S3: tbl_df/tbl/data.frame)
## $ Id: chr [1:3418] "001" "002" "003" "004" ...
## $ Q1: chr [1:3418] "SA" "SA" "A" "DK" ...
## $ Q2: chr [1:3418] "A" "SA" "SA" "SA" ...
## $ Q3: chr [1:3418] "A" "SA" "SD" "A" ...
## $ Q4: chr [1:3418] "SA" "SA" "SA" "A" ...
names(datos)
## [1] "Id" "Q1" "Q2" "Q3" "Q4"
head(datos)
## # A tibble: 6 × 5
## Id Q1 Q2 Q3 Q4
## <chr> <chr> <chr> <chr> <chr>
## 1 001 SA A A SA
## 2 002 SA SA SA SA
## 3 003 A SA SD SA
## 4 004 DK SA A A
## 5 005 A SA SD A
## 6 006 SA A A SA
# Limpiar nombres de variables
datos <- datos %>%
clean_names()
# Convertir variables a factor
datos_cat <- datos %>%
mutate(across(everything(), as.factor))
# Revisar valores perdidos
colSums(is.na(datos_cat))
## id q1 q2 q3 q4
## 0 0 0 0 0
# Crear carpeta de resultados
dir.create("resultados_mujer_trabajo", showWarnings = FALSE)
# ============================================================
# 3. SELECCIONAR VARIABLES DE PERCEPCIÓN
# ============================================================
# Se excluye id porque solo identifica a cada participante.
# El análisis se realiza únicamente con q1, q2, q3 y q4.
datos_percepcion <- datos_cat %>%
select(q1, q2, q3, q4)
# ============================================================
# 4. TABLAS UNIVARIADAS
# ============================================================
tablas_uni <- lapply(names(datos_percepcion), function(var){
datos_percepcion %>%
tabyl(!!sym(var)) %>%
adorn_totals("row") %>%
adorn_pct_formatting(digits = 1)
})
names(tablas_uni) <- names(datos_percepcion)
# Ver tabla de Q1
tablas_uni$q1
## q1 n percent
## A 476 13.9%
## DK 362 10.6%
## SA 2501 73.2%
## SD 79 2.3%
## Total 3418 100.0%
# Exportar tablas univariadas a Excel
wb_uni <- createWorkbook()
for(i in seq_along(tablas_uni)){
addWorksheet(wb_uni, paste0("Uni_", names(tablas_uni)[i]))
writeData(wb_uni, sheet = i, tablas_uni[[i]])
}
saveWorkbook(
wb_uni,
"resultados_mujer_trabajo/tablas_univariadas.xlsx",
overwrite = TRUE
)
# ============================================================
# 5. TABLAS BIVARIADAS
# ============================================================
pares <- combn(names(datos_percepcion), 2, simplify = FALSE)
crear_bivariada <- function(var1, var2){
datos_percepcion %>%
tabyl(!!sym(var1), !!sym(var2)) %>%
adorn_totals(c("row", "col")) %>%
adorn_percentages("row") %>%
adorn_pct_formatting(digits = 1)
}
tablas_bi <- lapply(pares, function(p){
crear_bivariada(p[1], p[2])
})
names(tablas_bi) <- sapply(pares, function(p){
paste0(p[1], "_vs_", p[2])
})
# Ver primera tabla bivariada
tablas_bi[[1]]
## q1 A DK SA Total
## A 38.2% 9.7% 52.1% 100.0%
## DK 37.6% 8.8% 53.6% 100.0%
## SA 38.3% 8.2% 53.5% 100.0%
## SD 29.1% 11.4% 59.5% 100.0%
## Total 38.0% 8.5% 53.5% 100.0%
# Exportar tablas bivariadas a Excel
wb_bi <- createWorkbook()
for(i in seq_along(tablas_bi)){
nombre_hoja <- substr(names(tablas_bi)[i], 1, 31)
addWorksheet(wb_bi, nombre_hoja)
writeData(wb_bi, sheet = i, tablas_bi[[i]])
}
saveWorkbook(
wb_bi,
"resultados_mujer_trabajo/tablas_bivariadas.xlsx",
overwrite = TRUE
)
# ============================================================
# 6. ANÁLISIS DE CORRESPONDENCIAS MÚLTIPLES MCA
# ============================================================
mca <- MCA(datos_percepcion, graph = FALSE)
# Resumen del MCA
summary(mca)
##
## Call:
## MCA(X = datos_percepcion, graph = FALSE)
##
##
## Eigenvalues
## Dim.1 Dim.2 Dim.3 Dim.4 Dim.5 Dim.6 Dim.7
## Variance 0.269 0.263 0.259 0.257 0.255 0.249 0.247
## % of var. 9.779 9.567 9.407 9.331 9.276 9.049 8.999
## Cumulative % of var. 9.779 19.345 28.752 38.084 47.359 56.408 65.408
## Dim.8 Dim.9 Dim.10 Dim.11
## Variance 0.242 0.240 0.236 0.233
## % of var. 8.812 8.736 8.585 8.460
## Cumulative % of var. 74.219 82.955 91.540 100.000
##
## Individuals (the 10 first)
## Dim.1 ctr cos2 Dim.2 ctr cos2 Dim.3 ctr cos2
## 1 | -0.085 0.001 0.009 | 0.212 0.005 0.053 | -0.180 0.004 0.038 |
## 2 | 0.376 0.015 0.056 | -0.454 0.023 0.082 | -0.506 0.029 0.103 |
## 3 | -0.424 0.020 0.059 | 0.108 0.001 0.004 | -0.384 0.017 0.049 |
## 4 | 0.162 0.003 0.008 | -0.092 0.001 0.003 | 0.849 0.081 0.226 |
## 5 | 0.251 0.007 0.018 | 0.147 0.002 0.006 | -0.386 0.017 0.042 |
## 6 | -0.085 0.001 0.009 | 0.212 0.005 0.053 | -0.180 0.004 0.038 |
## 7 | 0.148 0.002 0.019 | -0.290 0.009 0.072 | -0.098 0.001 0.008 |
## 8 | 0.589 0.038 0.255 | 0.251 0.007 0.046 | -0.182 0.004 0.024 |
## 9 | -0.526 0.030 0.423 | -0.329 0.012 0.165 | -0.095 0.001 0.014 |
## 10 | 0.817 0.073 0.248 | 0.087 0.001 0.003 | -0.591 0.039 0.130 |
##
## Categories (the 10 first)
## Dim.1 ctr cos2 v.test Dim.2 ctr cos2 v.test
## q1_A | -0.485 3.044 0.038 -11.400 | 0.953 12.017 0.147 22.406 |
## q1_DK | 0.089 0.078 0.001 1.788 | 0.271 0.741 0.009 5.460 |
## q1_SA | 0.062 0.259 0.010 5.960 | -0.134 1.241 0.049 -12.896 |
## q1_SD | 0.560 0.673 0.007 5.031 | -2.756 16.685 0.180 -24.782 |
## q2_A | 0.588 12.226 0.212 26.923 | 0.661 15.801 0.268 30.273 |
## q2_DK | -0.574 2.612 0.031 -10.246 | -0.141 0.160 0.002 -2.512 |
## q2_SA | -0.327 5.300 0.122 -20.457 | -0.448 10.187 0.230 -28.052 |
## q3_A | -0.444 11.196 0.309 -32.471 | 0.176 1.795 0.048 12.859 |
## q3_DK | 0.587 2.935 0.035 10.897 | -1.046 9.524 0.110 -19.416 |
## q3_SA | 1.427 20.986 0.254 29.454 | -0.080 0.067 0.001 -1.646 |
## Dim.3 ctr cos2 v.test
## q1_A -0.673 6.089 0.073 -15.816 |
## q1_DK 1.841 34.700 0.402 37.044 |
## q1_SA -0.084 0.504 0.019 -8.148 |
## q1_SD -1.712 6.548 0.069 -15.395 |
## q2_A -0.140 0.717 0.012 -6.394 |
## q2_DK 0.421 1.465 0.017 7.526 |
## q2_SA 0.032 0.053 0.001 2.005 |
## q3_A -0.077 0.354 0.009 -5.661 |
## q3_DK 1.780 28.026 0.319 33.028 |
## q3_SA -0.914 8.950 0.104 -18.866 |
##
## Categorical variables (eta2)
## Dim.1 Dim.2 Dim.3
## q1 | 0.044 0.323 0.495 |
## q2 | 0.217 0.275 0.023 |
## q3 | 0.396 0.120 0.387 |
## q4 | 0.419 0.334 0.129 |
# Eigenvalues: varianza explicada por dimensiones
eig <- get_eigenvalue(mca)
eig
## eigenvalue variance.percent cumulative.variance.percent
## Dim.1 0.2689142 9.778700 9.77870
## Dim.2 0.2630834 9.566668 19.34537
## Dim.3 0.2586942 9.407061 28.75243
## Dim.4 0.2566081 9.331205 38.08363
## Dim.5 0.2550840 9.275780 47.35941
## Dim.6 0.2488480 9.049018 56.40843
## Dim.7 0.2474809 8.999305 65.40774
## Dim.8 0.2423166 8.811512 74.21925
## Dim.9 0.2402315 8.735692 82.95494
## Dim.10 0.2360844 8.584888 91.53983
## Dim.11 0.2326547 8.460169 100.00000
write.xlsx(
eig,
"resultados_mujer_trabajo/eigenvalues_mca.xlsx"
)
# Coordenadas, contribuciones y calidad de representación
var_mca <- get_mca_var(mca)
coord_cat <- as.data.frame(var_mca$coord)
contrib_cat <- as.data.frame(var_mca$contrib)
cos2_cat <- as.data.frame(var_mca$cos2)
write.xlsx(
coord_cat,
"resultados_mujer_trabajo/coordenadas_categorias_mca.xlsx"
)
write.xlsx(
contrib_cat,
"resultados_mujer_trabajo/contribuciones_categorias_mca.xlsx"
)
write.xlsx(
cos2_cat,
"resultados_mujer_trabajo/calidad_representacion_cos2_mca.xlsx"
)
# ============================================================
# 7. GRÁFICOS DEL MCA
# ============================================================
# Gráfico de eigenvalues
grafico_eig <- fviz_screeplot(
mca,
addlabels = TRUE,
ylim = c(0, 50)
) +
ggtitle("Varianza explicada por dimensiones del MCA") +
theme_minimal()
grafico_eig
ggsave(
"resultados_mujer_trabajo/grafico_eigenvalues_mca.png",
plot = grafico_eig,
width = 8,
height = 5,
dpi = 300
)
# Mapa perceptual de categorías
grafico_var <- fviz_mca_var(
mca,
repel = TRUE,
col.var = "contrib",
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07")
) +
ggtitle("Mapa perceptual de categorías del MCA") +
theme_minimal()
grafico_var
ggsave(
"resultados_mujer_trabajo/mapa_perceptual_categorias_mca.png",
plot = grafico_var,
width = 9,
height = 7,
dpi = 300
)
exists("mca")
## [1] TRUE
# ============================================================
# Mapa de individuos sin etiquetas
# ============================================================
grafico_ind <- fviz_mca_ind(
mca,
label = "none",
geom = "point",
alpha.ind = 0.4
) +
ggtitle("Mapa de individuos del MCA") +
theme_minimal()
print(grafico_ind)
ggsave(
filename = "resultados_mujer_trabajo/mapa_individuos_mca.png",
plot = grafico_ind,
width = 9,
height = 7,
dpi = 300
)
# ============================================================
# Biplot del MCA mostrando categorías
# ============================================================
grafico_biplot <- fviz_mca_biplot(
mca,
label = "var",
repel = TRUE,
invisible = "ind",
ggtheme = theme_minimal()
) +
ggtitle("Biplot del MCA: categorías de percepción")
grafico_biplot
ggsave(
"resultados_mujer_trabajo/biplot_mca.png",
plot = grafico_biplot,
width = 9,
height = 7,
dpi = 300
)
# ============================================================
# 8. CONCLUSIONES Y RECOMENDACIONES
# ============================================================
# Conclusiones:
# 1. La base Woman Work Perc.xlsx contiene 3418 registros y cuatro preguntas
# de percepción: Q1, Q2, Q3 y Q4.
# 2. Las tablas univariadas permiten identificar la frecuencia y porcentaje
# de cada respuesta: SA, A, DK y SD.
# 3. Las tablas bivariadas permiten observar asociaciones descriptivas entre
# pares de preguntas, sin interpretar causalidad.
# 4. El MCA es adecuado porque las variables son categóricas y permite visualizar
# perfiles de respuesta mediante un mapa perceptual.
# 5. La variable Id no se interpreta porque solo identifica a cada participante.
# Recomendaciones:
# 1. Excluir Id de las tablas interpretativas y del MCA.
# 2. Analizar solo Q1, Q2, Q3 y Q4 como variables categóricas.
# 3. Recodificar las respuestas para facilitar la lectura:
# SA = totalmente de acuerdo
# A = de acuerdo
# DK = no sabe
# SD = totalmente en desacuerdo
# 4. Presentar en el informe una tabla univariada, una tabla bivariada
# representativa y el gráfico principal del MCA.
# 5. Usar lenguaje descriptivo: “se observa”, “se identifica” o “se asocia”,
# evitando afirmar relaciones causales.
# ============================================================
# CIERRE DEL ANÁLISIS
# ============================================================