How to reduce the time of df.describe() method execution for a spark dataframe with huge number of columns

Question

For one of my data validation script , I am doing df.describe() on the input dataframe for preparing data profile (where df is my input dataframe). For a particular dataframe of