I have read that you should use log transformations when the fluctuations on your data are increasing over time, but what do you do if the fluctuations level out over time?

A plot of the time series(1) and a log transform(2) are shown below:

Time Seires Data

Log Transform of Data


1 Answer 1


Consider what the purpose of the transformation is. Often the transformation is done to improve the normality of the data, or to help scale the data for use in time series analysis, model fitting, etc. In either case if that is what you want to do, probably a good idea in your case. You can try to compare the results with a normality plot before and after the log transformation.

As an aside, I recommend the Box-Cox transformation for time series work. Bit of a modification on the standard log transformation http://onlinestatbook.com/2/transformations/box-cox.html


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.