Using different scores can be problematic. It is important to ensure that those reading the chart are aware of the different scales.

In order for the different measures to fill the whole Y axis range, you could normalize the scores for each category, where the plotted score

Z = (Y – Ymin)/(Ymax-Ymin)

ranges from 0 to 1.

thanks

You need to define your parameters a, b, and c. Put them into a range of cells, and name the cells param_a, param_b, param_c. Then put the y values into a column, say, A2:A100. In B2 enter this formula:

=if(A2>=0,param_c^(param_a*A2)+param_b*A2-1,-param_c^(param_a*A2)+param_b*A2+1)

and fill down column B.

Thank you for all the excellent tutorials. I am working on a Parallel Coordinate graph for which the raw data has values between -1 and 24000 and am looking for an appropriate way to scale since I want to distinguish between 1 and 0, so far I’ve found

Bagwell, C Bruce. 2005. “Hyperlog-a flexible log-like transform for negative, zero, and positive valued data.” Cytometry. Part A : the journal of the International Society for Analytical Cytology 64:34-42. http://www.ncbi.nlm.nih.gov/pubmed/15700280.

but…am not sure how to implement it. Any ideas?

This is for my thesis and others have not been able to help so thank you very much!

That helps to show the amount of data for each line:

=B5+0.1*NORMSINV(RAND())