Plotly library makes interactive graphs. Using this library a function ddist has been written for visualization of data distribution of each variable within a dataset. This function may become quite handy during the exploration of any dataset.

Since this function uses plotly library, therefore you must install and load this library before calling the ddist function.

library(plotly)

ddist function takes a dataset (of type data.frame) as an input parameter and returns a list containing plotly plot object for each variable. See below:

#Function for generating plot objects for each variable within the dataset (using plotly library). For numeric and integer variables, a histogram plot is generated, while for others a barplot is generated.
#param: data.frame  
#returns: list
ddist <- function(dataset) {
  
  #create a list for holding the plot objects
  plots <- list(length(dataset))
  
  #iterate through each variable
  for(i in 1:length(dataset)) {
    
    #for numeric and integer variables plot histogram  
    if(is.numeric(dataset[,i]) || is.integer(dataset[,i])) {            
      plots[[i]] <- plot_ly(x=dataset[,i]) %>% add_histogram(name=names(dataset)[i])  
    } 
    #for remaining plot barplot
    else {
      tbl = table(dataset[,i])
      plots[[i]] <- plot_ly(x=names(tbl), y=tbl, name=names(dataset)[i], type='bar')
    }
  } 
  #return list of plots
  return(plots)
}

Now by using the above function we can easily explore the data distribution of each variable within any dataset. For example, the below code passes irisdataset to ddist function and then calls subplot function (of plotly library) to display the resulting plots on two rows:

#generate plotly plot objects
plots <- ddist(iris)

#display plots on two rows
subplot(plots, nrows=2)

Click to view interactive plots of univariate distributions of iris dataset

Similarly passing the diamonds dataset to ddist function results in:

Click to view interactive plots of univariate distributions of diamonds dataset