Installation | Syntax | Examples | Feedback | Change log
(03 Oct 2023)
This package provides the ability to draw joyplot or ridgeline plots in Stata. It is based on the Joyplot Guide that I released in October 2020.
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 package (v1.71) is available on SSC and can be installed as follows:
ssc install joyplot, replace
Or it can be installed from GitHub (v1.71):
net install joyplot, from("https://raw.githubusercontent.com/asjadnaqvi/stata-joyplot/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"
The syntax for the latest version is as follows:
joyplot y [x] [if] [in], by(variable)
[ overlap(num) bwidth(num) palette(str) alpha(num) offset(num) lines droplow normalize(local|global)
rescale offset(num) laboffset(num) lwidth(num) lcolor(str) ylabsize(num) ylabcolor(str) ylabposition(str)
yline ylcolor(str) ylwidth(str) ylpattern(str) xreverse yreverse peaks peaksize(num) n(num)
xtitle(str) ytitle(str) title(str) subtitle(str) xlabel(str)note(str) scheme(str) name(str) saving(str)
]
See the help file help joyplot
for details.
The most basic use is as follows:
joyplot y x, by(variable)
or
joyplot y, by(variable)
where y
is the variable we want to plot, and x
is usually the time dimension. The over
variable splits the data into different groupings that also determines the colors. The color schemes can be modified using the color()
option. Here any scheme from the colorpalettes
package can be used.
Set up the data:
clear
set scheme white_tableau
graph set window fontface "Arial Narrow"
use "https://github.com/asjadnaqvi/stata-joyplot/blob/main/data/OWID_data2.dta?raw=true", clear
We can generate basic graphs as follows:
joyplot new_cases date if date > 22460, by(country)
joyplot new_cases date if date > 22460, by(country) norm(local)
joyplot new_cases date if date > 22460, by(country) yline bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) alpha(100) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) lc(black) palette(white) alpha(100) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) lc(white) palette(black) alpha(50) lw(0.05) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) lines lw(0.2) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) lines lw(0.2) palette(black) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) lines lw(0.2) bwid(0.1) ylabpos(right) norm(local) offset(8)
joyplot new_cases date if date > 22460, by(country) lw(0.2) bwid(0.1) ylabpos(right) xsize(5) ysize(5) norm(local) offset(10)
joyplot new_cases date if date > 22460, by(country) bwid(0.1) yrev norm(local)
Here we reverse both axes, but it is highly advisable not to do so with the x-axis:
joyplot new_cases date if date > 22460, by(country) bwid(0.1) yrev xrev norm(local) yline laboff(-20)
joyplot new_cases date if date > 22460, by(country) bwid(0.1) norm(local)
joyplot new_cases date if date > 22460, by(country) bwid(0.1) overlap(15) norm(local)
joyplot new_cases date if date > 22460, by(country) bwid(0.1) overlap(12) lines norm(local)
joyplot new_cases date if date > 22460, by(country) bwid(0.1) off(-20) overlap(10) lw(none) norm(local)
We can also customize dates, increase the overlap of the layers, change the palette, and change the color intensity:
qui summ date if date > 22460
local xmin = r(min)
local xmax = r(max)
joyplot new_cases date if date > 22460, by(country) overlap(8) bwid(0.1) palette(CET C1) alpha(100) ///
lc(white) lw(0.2) xlabel(`xmin'(120)`xmax') off(-30) norm(local) ///
xtitle("Date") ytitle("Countries") ///
title("{fontface Arial Bold:My joyplot/ridgeline plot}") subtitle("Some more text here") ///
note("Some text here", size(vsmall))
Next we modify the scheme and make sure the colors are passed correctly. We use neon
from schemepack which has a black background:
qui summ date if date > 22460
local xmin = r(min)
local xmax = r(max)
joyplot new_cases date if date > 22460, by(country) overlap(8) bwid(0.1) palette(CET C1) alpha(90) ///
lc(black) lw(0.1) xlabel(`xmin'(120)`xmax') off(-30) norm(local) ///
ylabc(white) ///
xtitle("Date") ytitle("Countries") ///
title("{fontface Arial Bold:My joyplot}") subtitle("a subtitle here", color(white)) ///
note("Some text here", size(vsmall)) scheme(neon)
qui summ date if date > 22425
local xmin = r(min)
local xmax = r(max)
joyplot new_cases date if date > 22425, by(country) overlap(10) bwid(0.1) palette(black) alpha(100) norm(local) ///
lc(white) lw(0.2) xlabel(none) laboff(40) ///
ylabc(none) ///
xtitle("") ytitle("") ///
title("{fontface Arial Bold:The Joy Division look}") scheme(neon)
In v1.7, joyplot
can be replaced with ridgeline
, peaks
can be added to the ridges, and xline()
has been enabled:
ridgeline new_cases date if date > 22460, by(country) bwid(0.1) off(-20) overlap(8) peaks norm(local) xline(22700 23000)
Load the data that contains average USA state-level monthly temperatures for the period 1991-2020:
use "https://github.com/asjadnaqvi/stata-joyplot/blob/main/data/us_meantemp.dta?raw=true", clear
lab de month 1 "Jan" 2 "Feb" 3 "Mar" 4 "Apr" 5 "May" 6 "Jun" 7 "Jul" 8 "Aug" 9 "Sep" 10 "Oct" 11 "Nov" 12 "Dec", replace
lab val month month
joyplot meantemp, by(month)
joyplot meantemp, by(month) yline bwid(1.2)
joyplot meantemp, by(month) yline yrev bwid(1.2)
joyplot meantemp, by(month) yline ylw(0.2) ylc(blue) ylp(dot) ylabpos(right) bwid(1.2)
joyplot meantemp, by(month) bwid(1.5) ylabs(3) overlap(3) yline yrev palette(CET C6) ///
xlabel(-20(10)30) ///
ytitle("Month") xtitle("degrees Centigrade") ///
title("Mean average temperature in the USA") subtitle("2009-2020 average") ///
note("Source: World Bank Climate Change Knowledge Portal (CCKP).", size(vsmall)) ///
xsize(4) ysize(5)
qui summ meantemp
local xmin = r(min)
local xmax = r(max)
joyplot meantemp, by(month) bwid(1.5) ylabs(3) overlap(3) yline palette(scico corkO) alpha(100) ///
ytitle("Month") xtitle("degrees Centigrade") xlabel(`xmin'(5)`xmax') ///
title("Mean average temperature in the USA") subtitle("2009-2020 average") ///
note("Source: World Bank Climate Change Knowledge Portal (CCKP).", size(vsmall)) ///
xsize(3) ysize(5)
ridgeline meantemp, by(month) bwid(1.5) ylabs(3) overlap(3) lc(black) yline yrev palette(CET C6) ///
xlabel(-20(10)30) ///
ytitle("Month") xtitle("degrees Centigrade") ///
title("Mean average temperature in the USA") subtitle("2009-2020 average") ///
note("Source: World Bank Climate Change Knowledge Portal (CCKP).", size(vsmall)) ///
xsize(4) ysize(5) xline(-10 0 10) peak peaksize(0.4) // saving(test1)
Load a dummy data set
use "https://github.com/asjadnaqvi/stata-joyplot/blob/main/data/rescale_test.dta?raw=true", clear
use rescale_test, clear
drop if inlist(country, "Cambodia", "Myanmar", "Lao PDR")
tab country
Do a vanilla joyplot:
joyplot socMob year, by(country) ///
lc(white) lw(0.2) xlabel(1990(5)2020)
Get rid of the overlaps. Here overlap(1)
gives each country it's own box:
joyplot socMob year, by(country) overlap(1) rescale ///
lc(white) lw(0.2) xlabel(1990(5)2020)
Let's make a country unusable:
drop if year < 2015 & country=="Vietnam"
joyplot socMob year, by(country) overlap(1) ///
lc(white) lw(0.2) off(-2) xlabel(1990(5)2020)
The above code will produce an error message, highlighting where the errors exist. Let's throw these out using the droplow
option:
joyplot socMob year, by(country) overlap(1) droplow ///
lc(white) lw(0.2) off(-2) xlabel(1990(5)2020)
And we rescale the data further to the minimum and maximum values:
joyplot socMob year, by(country) overlap(1) droplow rescale ///
lc(white) lw(0.2) off(-2) xlabel(1990(5)2020)
Please open an issue to report errors, feature enhancements, and/or other requests.
v1.71 (03 Oct 2023)
- Fixed a bug where locals were passing incorrectly (reported by osnofas).
- Fixed a bug for the
lines
option in single density joyplots (reported by osnofas). - Added the
n()
option to evalute densities using a higher number of points (requested by osnofas). - Improved the error messages in the help file to better reflect the issues.
v1.7 (14 Jul 2023)
peaks
andpeaksize()
option added to mark highest point on ridges and modify their size. This option is currently beta.xline()
option added to allow users to plot reference lines on the x-axis (requested by Glenn Harrison).saving()
option added (requested by Glenn Harrison).joyplot
is now fully mirrored by theridgeline
command. Several users requested this.- Several bug fixes and better arrangement of the helpfile.
v1.62 (28 May 2023)
- Changed
over()
toby()
to align it with other similar packages:streamplot
,bumpline
,bumparea
. - Added
laboffset()
to displace labels in pixels. - Added
offset()
to extend the x-axis. - Minor fixes to code, help file, and updates to defaults.
v1.61 (01 Mar 2023)
- Fixed a bug where y-labels were reversed.
- Fixed a bug where
norm()
was not passing correctly to densities.
v1.6 (05 Nov 2022)
color
renamed topalette
to align it with the other packages- Errors checks now halt the program if
over
groups have fewer than 10 observations. This is to prevent density functions from crashing with low data points. - New option
norm(global)
ornorm(local)
allows users to draw lines normalized to global or local group-wise maximum values. - New option
droplow
drops the groups that do not need the above observation criteria. - New option
rescale
rescales the data to start from 0. This allows users to rebase values that start at very high or negative values.
v1.5 (03 Sep 2022)
- Code clean up.
- Default values updated.
- Error checks added.
v1.42 (22 Jun 2022)
- Values were wrongly assigned on the y-axis if the over variable was numeric (reported by Barry Burden).
- Fixes to "over" variable. It can now take on numeric, string, and labeled numeric values.
v1.41 (20 Jun 2022)
- Fix to labmask installation
- over now takes numerical values
v1.4 (26 Apr 2022)
- option to reverse x and y axes added
- option save graph names added
- optimization to x-axis scales
- various bug fixes
v1.3 (24 Apr 2022)
- option to add y-axis lines.
- the syntax is now allows for variable level densities.
- label placement optimized.
v1.21 (15 Apr 2022)
- y label position option added. It can now be position on the left or right.
- xsize and ysize options added to change the graph dimensions.
v1.2 (14 Apr 2022)
- x-axis angle option added.
- Global normalization option added.
- Draw lines only option added.
- Major fixes to syntax, defaults values, and how lines and areas are drawn.
v1.1 (08 Apr 2022)
- Public release. Several options and features added.
v1.0 (13 Dec 2021)
- Beta version.