logarithmic binning data in matlab

I have a dataset, and I have been told to bin the data using logarithmic binning.

How would I go about logarithmic binning this dataset?

Answer 1

[John D'Errico]

It is not clear what you mean by logarithmic binning.

Looking at your data however, tells me that you were told to do something that makes no sense at all.

min(Q)
ans =
-4.77253376114796

max(Q)
ans =
47.6051494981383

mode(Q)
ans =
     0
     
sum(Q == 0)
ans =
       22891
       
numel(Q)
ans =
       57158

So your data goes from negative to positive values, but there is a HUGE spike at EXACTLY 0. Literally half of your data is EXACTLY zero. Not just close. But EXACTLY so.

histogram(Q,1000)

Sorry. In no circumstances would any sort of traditional logarithmic binning like that make sense. And that HUGE spike at 0 makes things almost impossible to work with.

Could you choose bins that grow in size, as you move away from zero, in BOTH directions? Well, yes.

I would recommend you return to whoever told you to do this, and show them your actual data. Then ask them SPECIFICALLY what they think you should do. Don't just let them tell you to do something by some vague name. Have them tell you explicitly what they think is appropriate here.