When calculating a histogram we do data binning, or group a number of more or less continuous values into a smaller number of "bins". But in bucket sort we set up buckets and assign a bucket to each value of some collection, according to its value. Although for a different purpose, the methods seem similar to me.

In statistics, what is the difference between bin/binning, and bucket/bucketing? For example, if I have a set of $N$ observations sampling $y = f(x_1, x_2, x_3)$, and I want to estimate how some statistic varies over my domain, one way is to put my $N$ observations in bins/buckets according to $(x_1, x_2, x_3, y)$ and calculate my statistic within each bin/bucket. An example could be the mean and standard deviation of temperature as a function of latitude, longitude, and altitude.

In this example, is there a standard term for those entities I divide my data in? Are those more correctly called bins, or buckets? Or does it really not matter?


1 Answer 1


A very good question, and a question that I myself had because I have heard these called buckets, groups, groupings, categories, categorical variables, discrete variables, and bins as I have changed disciplines. In general, use the language that the end-users of your analysis are most comfortable using - in a sense, speak their language (or force them to use yours! ha). There is no wrong answer here, other than a countless number of statisticians that would say that you shouldn't be grouping your variables into bins/buckets without a very good reason (or ever!) as you are spending degrees of freedom, making arbitrary cutoffs to create your buckets/bins, and losing information that was provided by your once valuable, continuous variables.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.