Options to plot yearly data for five years in a single graph

[Chris Boulis]

Hi All.

I'd like to plot two different graphs using the same data.

(1) yearly data over five waves in a single graph:

tw (line totasset_ecpc agegrp if wave == 2 & inrange(agegrp, 30, 80))  ///
    (line totasset_ecpc agegrp if wave == 6 & inrange(agegrp, 30, 80))  ///
    (line totasset_ecpc agegrp if wave == 10 & inrange(agegrp, 30, 80))  ///
    (line totasset_ecpc agegrp if wave == 14 & inrange(agegrp, 30, 80))  ///
    (line totasset_ecpc agegrp if wave == 18 & inrange(agegrp, 30, 80)),  ///
title("Share of Assets - ECPC") name(ecpc, replace) ///
xla(30(10)80) xtitle("age") ytitle(" Proportion") yla(.3(.1).8) ///
legend(order(1 "Wave 2" 2 "Wave 6" 3 "Wave 10" 4 "Wave 14" 5 "Wave 18")) ///
legend(region(lstyle(none)) size(small) col(3) row(1) pos(6) rowgap(1) span)

(2) combine data for all five waves into a single plot:

tw (scatter totasset_ecpc agegrp if inrange(agegrp, 30, 80)), title("Share of Assets - ECPC (all waves)") ///
xla(30(10)80) xtitle("age") ytitle(" Share of assets") yla(.3(.1).8) name(ecpcall, replace)

While trends are observable in each of these graphs, they are cluttered. I'd appreciate options that increase their clarity.

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte wave float(totasset_ecpc agegrp)
 2  .7590239 20
 6         . 20
10  .8799195 20
14  .2183252 20
18         . 20
 2  .2032862 25
 6    .44044 25
10  .4350298 25
14  .4296866 25
18  .8124846 25
 2  .5380963 30
 6  .3945256 30
10  .4283356 30
14  .5113511 30
18  .3327902 30
 2  .5441144 35
 6  .4536623 35
10  .4263607 35
14   .397131 35
18 .48152775 35
 2   .531788 40
 6  .6135195 40
10  .4566517 40
14 .51987433 40
18 .39560035 40
 2   .501982 45
 6 .50936675 45
10  .5787155 45
14 .53960353 45
18  .4934241 45
 2  .5184313 50
 6  .4877012 50
10 .51347977 50
14   .542489 50
18 .55537874 50
 2  .5851486 55
 6   .511164 55
10  .4935571 55
14  .4886676 55
18  .5297553 55
 2  .5776826 60
 6  .5293519 60
10  .5640008 60
14  .4978073 60
18  .4536004 60
 2 .55672157 65
 6  .6667274 65
10  .5255127 65
14  .5557258 65
18  .4618697 65
 2  .6269194 70
 6   .532701 70
10  .5955692 70
14  .5492495 70
18 .55500907 70
 2   .505511 75
 6  .6363178 75
10  .5910302 75
14  .4802268 75
18   .542687 75
 2 .27799776 80
 6  .4065089 80
10  .4931756 80
14  .6125295 80
18  .5147132 80
end

Stata v.15.1. Using panel data.

[Maarten Buis]

you could look at Nick Cox's fabplot, type in Stata ssc desc fabplot

[Nick Cox]

https://www.statalist.org/forums/for...e-on-ssc/page3 is perhaps where to start. A paper on front-and-back plots will appear in Stata Journal 21(2).

[Chris Boulis]

Thank you Maarten Buis. I've installed fabplot and reading Nick's link as I write this. Thanks Nick Cox for the link to subsetplot & fabplot. I read Cox, N.J. 2010. Graphing subsets. Stata Journal 10: 670-681 and found a really nice way to display data by wave using code on pp. 675-676. That said, subsetplot gave close to the same with a lot less code. I also really like fabplot (though is there a way to add age to graphs 2 & 6?).

fabplot line totasset_ecpc agegrp, by(wave) ///
yti("") xti("") yla(.1(.1).8, ang(h)) xla(20(10)80) legend(off) ///
front(connect) frontopts(mc(blue) lc(blue)) name(G4, replace)

I want to observe if there are trends in the data (ratios - proportion of assets of one group compared to another) that is, if data points are clearly higher/lower than 0.5? Finally, is it possible to achieve this on a smaller footprint (a single graph), given the 'spaghetti' problem?