Installation | Syntax | Examples | Feedback | Change log

---

---

ternary v1.3

(07 May 2025)

This package provides the ability to draw tri-variate plots in Stata.

Installation

The package can be installed via SSC or GitHub. The GitHub version, might be more recent due to bug fixes, feature updates etc, and may contain syntax improvements and changes in default values. See version numbers below. Eventually the GitHub version is published on SSC.

The SSC version (v1.2):

ssc install ternary, replace

Or it can be installed from GitHub (v1.3):

net install ternary, from("https://raw.githubusercontent.com/asjadnaqvi/stata-ternary/main/installation/") replace

The palettes package is required to run this command:

ssc install palettes, replace
ssc install colrspace, replace

Even if you have the package installed, make sure that it is updated ado update, update.

If you want to make a clean figure, then it is advisable to load a clean scheme. These are several available and I personally use the following:

ssc install schemepack, replace
set scheme white_tableau  

You can also push the scheme directly into the graph using the scheme(schemename) option. See the help file for details or the example below.

I also prefer narrow fonts in figures with long labels. You can change this as follows:

graph set window fontface "Arial Narrow"

Syntax

The syntax for the latest version is as follows:

ternary varL varR varB [if] [in], [ by(var) cuts(num) zoom pad(num) normalize(1|100) nofill points lines labels colorL(str) colorR(str)
	colorB(str) lwidth(str) msymbol(str) msize(str) mcolor(str) mlcolor(str) mlwidth(str) labcolor(str) ticksize(str) mlabel(var)
	mlabsize(str) mlabcolor(str) mlabposition(str) scale format(fmt) palette(str) legend(options) * ]

See the help file help ternary for details.

The most basic use is as follows:

ternary varL varR varB

representing left, right and bottom variables respectively.

Citation guidelines

Software packages take countless hours of programming, testing, and bug fixing. If you use this package, then a citation would be highly appreciated.

The SSC citation is recommended. Please note that the GitHub version might be newer than the SSC version.

Examples

Users can either copy the files locally

foreach x in NUTS2_edu NUTS2_tourstay NUTS3_gva NUTS3_pop {
	copy "https://github.com/asjadnaqvi/stata-ternary/raw/main/data/`x'.dta" "`x'.dta", replace
}

or load them directly from the server:

use "https://github.com/asjadnaqvi/stata-ternary/blob/main/data/NUTS3_pop.dta?raw=true", clear

Test a basic figure

ternary y15prop y64prop y99prop

Rearrange the variables and remove the fill

ternary  y99prop y15prop y64prop, nofill

Test ternary with fractions:

ternary y99prop y15prop y64prop, norm(1)

Colors

ternary  y99prop y15prop y64prop

ternary  y99prop y15prop y64prop, lcolor(black)

Cuts

ternary y99prop y15prop y64prop, cuts(3)

ternary y99prop y15prop y64prop, cuts(10)

ternary y99prop y15prop y64prop, cuts(6)

Zoom

ternary y99prop y15prop y64prop, cuts(6) zoom

ternary y99prop y15prop y64prop, cuts(6) zoom msize(2.5) mc(black%60) mlc(white)

ternary y99prop y15prop y64prop, cuts(8) zoom msize(2) mc(black%70)  mlc(white) lw(0.08) labcolor(gs6) 

labels (v1.1)

ternary y99prop y15prop y64prop, cuts(8) zoom lw(0.08) labcolor(gs6) mlabel(NUTS_ID)

ternary y99prop y15prop y64prop, cuts(8) zoom lw(0.08) labcolor(gs6) ///
	mlabel(NUTS_ID)  mlabpos(12)

gen NUTS_ID2 = NUTS_ID if y99prop > 30	

ternary y99prop y15prop y64prop, cuts(8) zoom lw(0.08) labcolor(gs6) ///
	mlabel(NUTS_ID2)  mlabpos(12)

Customize colors

ternary y99prop y15prop y64prop, zoom cuts(6) mcolor(white%60) ///
	colorB(lime) 

ternary y99prop y15prop y64prop, zoom cuts(6) ///
	colorB(red) colorL(green) colorR(blue)

ternary y99prop y15prop y64prop, zoom cuts(4) ///
	colorB(red) colorL(green) colorR(blue) mcolor(white%60) mlc(black) msize(3)	

other options

ternary y99prop y15prop y64prop, nofill points zoom cuts(6)

ternary y99prop y15prop y64prop, points zoom cuts(6) mlcolor(white)

ternary y99prop y15prop y64prop, nofill points zoom cuts(6) lw(0.05) msize(2.5) mlc(white)

ternary y99prop y15prop y64prop, nofill lines labels zoom cuts(6) msize(2.5) points malpha(80) mlc(white)

ternary y99prop y15prop y64prop, nofill lines labels zoom cuts(6) msize(1.5) 

ternary y99prop y15prop y64prop, nofill lines labels zoom cuts(6) msize(1.5) mlc(black) mc(white%60)

ternary y99prop y15prop y64prop, nofill lines labels zoom cuts(6) msize(2) mlc(black) mc(white%80) lw(0.3) msym(triangle)

v1.2

Plot the raw data and scale the markers to the total population:

ternary y_Y_GE65 y_Y15_64 y_Y_LT15, zoom lw(0.08) scale  norm(100) mcolor(white%70) msize(0.8)

gen ctry=substr(NUTS_ID, 1, 2)

ternary y_Y_GE65 y_Y15_64 y_Y_LT15 if ctry=="DE", zoom lw(0.08) scale  norm(100) mcolor(white%60) msize(0.8) pad(2)

v1.3 (by categories)

Generate a region variable for EU countries:

*** taken from https://en.wikipedia.org/wiki/EuroVoc

gen region = .

replace region = 1 if inlist(ctry, "AT", "BE", "FR", "DE", "IE" , "LI", "LU", "MK")
replace region = 1 if inlist(ctry, "NL", "CH", "UK")

replace region = 2 if inlist(ctry, "DK", "IS", "EE", "FI", "LV", "LT", "SI", "NO")

replace region = 3 if inlist(ctry, "CY", "EL", "IT", "MT", "PT" , "ES", "TR")

replace region = 4 if inlist(ctry, "AL", "CZ", "HR", "HU", "PL", "RO", "RS")
replace region = 4 if inlist(ctry, "SE", "SK", "MK", "ME", "BG")

lab de region 4 "Central and Eastern Europe" 2 "Northern Europe" 3 "Southern Europe" 1 "Western Europe", replace
lab val region region

ternary y_Y_GE65 y_Y15_64 y_Y_LT15 , by(region) cuts(4)  zoom lw(0.08) norm(100) mcolor(white%70) msize(1.8) pad(2)

ternary y_Y_GE65 y_Y15_64 y_Y_LT15 , by(region) cuts(4) nofill zoom lw(0.08) norm(100) mcolor(white%70) msize(1.2) pad(2)

ternary y_Y_GE65 y_Y15_64 y_Y_LT15 , by(region) cuts(2) zoom lw(0.08) norm(100) msymbol(triangle circle square diamond) msize(0.2) pad(2) nofill scale legend(size(2) pos(6) rows(2))

ternary y_Y_GE65 y_Y15_64 y_Y_LT15, by(region) cuts(2) nofill malpha(100) zoom lw(0.08) norm(100) msymbol(circle) msize(1) pad(2) palette(burd) mlcolor(white) mlwidth(0.06)

Feedback

Please open an issue to report errors, feature enhancements, and/or other requests.

Change log

v1.3 (07 May 2025)

• Option by() added to allow markers to vary by categories.
• Option msymbol() now accepts lists. If there are fewer markers defined than the by categories, then the last marker will be used for the remaining.
• Option palette() added to customize marker colors.
• Option legend() added to control how by() legends are drawn.
• Various bug fixes.

v1.2 (12 Mar 2025)

• Option fill is now the default and has been removed. Instead nofill has been added to remove the colors. This aligns the package to also how it was intended to be drawn by default.
• Option norm(100) show now work properly.
• Option scale has been added to scale the markers to the total of the raw data. In order to use this correctly, specify the raw data and select norm(100) or norm(1).
• Option zoom should now result in tighter bounds.
• Option pad() has been added to extended the zoom bounds. Default is pad(5) fof 5%.

v1.1 (12 Sep 2024)

• Added norm() option which can be used to normalize the data to 1 (norm(1)) or 100 (norm(100)).
• Added marker label options: mlabel(), mlabcolor(), mlabposition(), mlabsize().
• Better zoom option.
• Minor code cleanups.

v1.0 (28 Aug 2024)

• First release.