2
$\begingroup$

I am trying to compare multiple columns of data with different units graphically. If I z-score the whole dataset am I introducing any error? Can I safely compare from column to column?

Eg

Pb(mg) CO(kg) N(?)
42.         6543.   123
49.         7989.    456
54.          4577.   222
$\endgroup$
  • $\begingroup$ By comparing relationships are you implying correlation between two variables? $\endgroup$ – Arun Jose Nov 4 '16 at 11:15
  • $\begingroup$ No, I simply want to know how I can compare mg of lead (Pb) to kg of carbon monoxide (CO) on the same Y axis. Will Z-scoring keep me from having to convert the units? $\endgroup$ – Rilcon42 Nov 4 '16 at 14:22
0
$\begingroup$

Z-score will preserve the shape of the distribution, and you will be able to compare them, such as noticed if one or more skewed than the others.

Here are box plots of five variables using raw values and using z-scores.

enter image description here

enter image description here

Of course, with the z-scores you don't need to repeat the Y axis as I have, but I did want to point out the idea of parallel coordinates here as an alternative to z-scores. That is you leave the data as raw values and show the native axes side-by-side (parallel to each other).

It works for any univariate display. Here are histograms.

enter image description here

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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