exponential scale for plot axes

[Sharad Kumar]

I am trying to plot distributions of article citations. Due to the skewness of citations, I want the axes to be measured in an exponential scale. Does stata provide this option? See below image for an illustration of the type of scaling I am looking for (image copied from Radicchi, Fortunato & Catellano. 2008. PNAS).

Below is a MWE for the code I am using to currently plot citations. I am looking for an option that I could specify at the end of this MWE to achieve the exponential scale as in Radicchi et al. Currently the scale is in levels and the plot looks squished.

twoway (kdensity citations if Engineering==1, lcolor(gs7) lwidth(medthick)) || ///
       (kdensity citations if Biology==1, lcolor(gs7) lwidth(medthick) lpattern(shortdash))

[Nick Cox]

I'd call those logarithmic scales.

I don't think many people in these fields use kernel density estimation at all, and you're getting to have to work hard to get similar results with that method.

It's not even clear to me that the vertical axis shows probability density: it looks like probability.

They just have histograms in effect shown as segmented curves, with some adjustment, or so one hopes, for unequal bin width.

Your idea of a MWE differs from mine. I haven't tried to find or read the paper.