Data Explorar

Teoría

La líbreria DataExplorer es la mas conocida en el mundo análitico exploratorio. Es muy simple de usar y muy poderosa, pues ofrece como salida un informe con mucha información.

La función para crear el informe es create_report, y para ver cada gráfica de forma individual, las funciones son:

introduce() plot_intro() plot_boxplot() plot_missing() plot_histogram plot_bar() plot_correlation()

Instalar paquetes y llamar librerías

#install.packages("DataExplorer")
library(DataExplorer)

## Warning: package 'DataExplorer' was built under R version 4.3.1

#install.packages("nycflights13")
library(nycflights13)

Contexto

El paquete nycflights13 contiene información sobre todos los vuelos que partieron desde Nueva York (EWR, JFK y LGA) a destinos en los Estados Unidos en 2013. FUeron 336,776 vuelos en total.

Las tablas de este paquete y sus relaciones son las siguientes:

Crear base de datos

flights <- flights
weather <- weather
plnes <- planes
airports <- airports
airlines <- airlines
df <- merge(flights,airlines,by = "carrier")
df <- merge(df, planes, by = "tailnum")

introduce(df)

##     rows columns discrete_columns continuous_columns all_missing_columns
## 1 284170      28               10                 18                   0
##   total_missing_values complete_rows total_observations memory_usage
## 1               311768           920            7956760     50225296

plot_intro(df)

plot_boxplot(df, by = "carrier")

## Warning: Removed 23255 rows containing non-finite values (`stat_boxplot()`).

## Warning: Removed 288513 rows containing non-finite values (`stat_boxplot()`).

plot_missing(df)

plot_histogram

## function (data, binary_as_factor = TRUE, geom_histogram_args = list(bins = 30L), 
##     scale_x = "continuous", title = NULL, ggtheme = theme_gray(), 
##     theme_config = list(), nrow = 4L, ncol = 4L, parallel = FALSE) 
## {
##     variable <- value <- NULL
##     if (!is.data.table(data)) 
##         data <- data.table(data)
##     split_data <- split_columns(data, binary_as_factor = binary_as_factor)
##     if (split_data$num_continuous == 0) 
##         stop("No continuous features found!")
##     continuous <- split_data$continuous
##     feature_names <- names(continuous)
##     dt <- suppressWarnings(melt.data.table(continuous, measure.vars = feature_names, 
##         variable.factor = FALSE))
##     layout <- .getPageLayout(nrow, ncol, ncol(continuous))
##     plot_list <- .lapply(parallel = parallel, X = layout, FUN = function(x) {
##         ggplot(dt[variable %in% feature_names[x]], aes(x = value)) + 
##             do.call("geom_histogram", c(na.rm = TRUE, geom_histogram_args)) + 
##             do.call(paste0("scale_x_", scale_x), list()) + ylab("Frequency")
##     })
##     class(plot_list) <- c("multiple", class(plot_list))
##     plotDataExplorer(plot_obj = plot_list, page_layout = layout, 
##         title = title, ggtheme = ggtheme, theme_config = theme_config, 
##         facet_wrap_args = list(facet = ~variable, nrow = nrow, 
##             ncol = ncol, scales = "free"))
## }
## <bytecode: 0x119b45a78>
## <environment: namespace:DataExplorer>

plot_bar(df)

## 4 columns ignored with more than 50 categories.
## tailnum: 3322 categories
## dest: 104 categories
## time_hour: 6934 categories
## model: 127 categories

plot_correlation(df)

## 5 features with more than 20 categories ignored!
## tailnum: 3322 categories
## dest: 104 categories
## time_hour: 6934 categories
## manufacturer: 35 categories
## model: 127 categories

## Warning in cor(x = structure(list(year.x = c(2013L, 2013L, 2013L, 2013L, : the
## standard deviation is zero