I started with a data frame with about 20\'000 observations and around 200 columns. Since its a survey I divided the columns up in around 6 sections. I did some calculations lik