Overlay a two-way graph and an hisogram from two distinct datasets

[Nicolas Charette]

Hello,

I’m using Stata BE/17.0. I want to overlay a twoway graph from data A over an histogram from data B (or vice-versa). I want the final figure to have two y axis. My intuition is that I would need to have my info in only one dataset. Is it feasible with two separate datasets using a saved graph/histogram? Any suggestions?

Data A (twoway graph):

* Example generated by -dataex-. For more info, type help dataex
clear
input float duree double(estimate min95 max95)
 30 -.39709141041976465 -1.7826537251028372  .9884709042633077
 60 -.39709141041976465 -1.7826537251028372  .9884709042633077
 90 -.45052648862218997 -1.8140828455056424  .9130298682612623
120  -.5234470219012863 -1.7725581317823662  .7256640879797935
150  -.5234470219012863 -1.7725581317823662  .7256640879797935
180  -.5646014010734139 -1.9337486345279202  .8045458323810923
210   -.790618265194683 -2.1615262120481495  .5802896816587834
240  -.7445110715021096 -1.9676471592732083 .47862501626898923
270  -.9396592958036535 -2.1593530138289143 .28003442222160746
end

Code for the two-way graph, where y axis is an estimate of a program return in a school subject and x axis values are numbers of days of participation to the program:

twoway (scatter estimate duree,legend(region(col(white))) yscale(titlegap(2)) xscale(range(0 300) titlegap(2)) msymbol(square) msize(small)) (rcap min95 max95 duree, lcolor(blue)), xlabel(0(30)300, valuelabels angle(0) labsize(small))  xtitle("Number of days") ytitle("Grade variation (%)") ysca(titlegap(2)) ylab(, labs(small)) xsca(titlegap(2)) yline(0, lcolor(red)) graphregion(color(white))

Dataset B (histogram):

* Example generated by -dataex-. For more info, type help dataex
clear
input float(id jours)
  6 214
 35 181
 41 225
 51 207
 58 206
 63 184
 68 227
 77 150
 80 165
102  64
131 241
136 261
140 177
162 192
163  76
166 236
176 157
185 106
209  87
213 269
237  83
266 239
277  10
359 191
360 191
373  97
end

Code for the histogram, where y axis is a proportion in percentage and x axis values are numbers of days of participation to the program:

histogram jours if jours>0 & jours<270, xscale(range(0 300)) percent fcolor(bluishgray) ytitle(Proportion (%)) title(Proportion of students in the program by days of participation) xlabel(0(30)300) bin (10) name(jours, replace)

Thank you for your help,

Nicolas Charette

[Nicolas Charette]

Hello,

I will try to reformulate. I have modified the data. I think what I wanted to do was impossible. TO THE STATALIST VETERANS: Should I create a new post?

I created a twoway scatter diagram that shows the estimated coefficients of the impact of a stay-in-school program on the students grade, depending on the number of days they participated in the program. For each estimate, the minimum of days required to be included in the treated group for the regression goes up by 30 days, so as the number of days goes up, the number of treated students in the treated group (variable “treated”) goes down. The variable “proportion” shows the proportion of the initial 315 students that are considered treated for that number of days. I want to add the number of student included in the treated group for each estimate, but I’m having problems.

* Example generated by -dataex-. For more info, type help dataex
clear
input float(days treated proportion) double(estimate min95 max95)
  30 314  .9968254 .046494984169676644 -1.9137120599203097  2.006702028259663
  60 312  .9904762 .046494984169676644 -1.9137120599203097  2.006702028259663
  90 307  .9746032   .7679641564922828 -1.2171593010667443   2.75308761405131
 120 304  .9650794   .7053769464142945 -1.3752948607208155 2.7860487535494043
 150 303  .9619048   .7053769464142945 -1.3752948607208155 2.7860487535494043
 180 299  .9492064   .7496901996016438  -1.094470087186972 2.5938504863902594
 210 288  .9142857   .8744209239132374   -.937688168573908 2.6865300164003827
 240 283  .8984127  1.0652281440416944  -.6800340685545461  2.810490356637935
 270 277  .8793651  1.2252341610868356  -.5849420620836219 3.0354103842572933
 300 269  .8539683   1.622480267793473 -.09256744557846797 3.3375279811654144
 330 266  .8444445  1.6464548726504031  -.0454302588594071  3.338340004160213
 360 262   .831746  1.7303964507057465 .031639609180645456  3.429153292230848
 390 252        .8  2.1070893356168483   .3949298994637567   3.81924877176994
 420 244  .7746032  2.7412028929879013   .8906725894595471  4.591733196516255
 450 238  .7555556  2.3624295839629177   .7458820651322966 3.9789771027935386
 480 233  .7396826  2.7401356084756063  1.2453542045037043  4.234917012447508
 510 229  .7269841  3.2365693377283065  1.6689275614528654  4.804211114003747
 540 225  .7142857  2.9633489603477297  1.4576143258043381  4.469083594891121
 570 222  .7047619  3.0624508868160554  1.5549986274808616  4.569903146151249
 600 216  .6857143  2.9580537025950657   1.500419624515906  4.415687780674226
 630 214  .6793651   2.802549173921975   1.362511785432553  4.242586562411397
 660 205  .6507937  2.8552935302961844  1.6429667871494913  4.067620273442877
 690 199  .6317461    2.60091867213188  1.4055934379393151  3.796243906324445
 720 196  .6222222   2.862556692980679  1.7134561672942563  4.011657218667102
 750 190  .6031746   3.020192408941428  1.7886077570320789 4.2517770608507774
 780 180  .5714286   3.270420375869851  2.1572063649020325  4.383634386837669
 810 179   .568254  3.2798848434042385  2.1580191162282016  4.401750570580275
 840 178  .5650794   3.334238577065541   2.190461802392833  4.478015351738249
 870 174   .552381  3.2029810659708073  1.9022880866712197  4.503674045270395
 900 171 .54285717  3.2249530746133077  1.7909727176552197  4.658933431571396
 930 169 .53650796  2.9694181687471892  1.4199996720686934  4.518836665425685
 960 165 .52380955  2.9693774549328418  1.4562584482583165  4.482496461607367
 990 160 .50793654  3.1921786652969804  1.5952422399213306   4.78911509067263
1020 159 .50476193  3.1921786652969804  1.5952422399213306   4.78911509067263
1050 153  .4857143  3.0669962431122233  1.3994037984757741  4.734588687748673
1080 145  .4603175  3.3090450937010547  1.6462209310607643  4.971869256341345
1110 142 .45079365   3.728517834601178   2.228913350223377  5.228122318978979
1140 140 .44444445  3.7170518592235777  2.3227947841369128  5.111308934310243
1170 137  .4349206  3.5958287913156295  1.9852390918627096   5.20641849076855
1200 136   .431746  3.6005460550636914  1.9264560953962446  5.274636014731138
1230 136   .431746  3.6005460550636914  1.9264560953962446  5.274636014731138
1260 134  .4253968   3.720102331499618  1.9221698018836166   5.51803486111562
1290 132  .4190476   3.948177458153095  2.0506693790139403  5.845685537292249
1320 131   .415873  3.6746230141674796   1.668615296676586  5.680630731658374
1350 129  .4095238  3.6848180255306016   2.035406144410557  5.334229906650647
1380 117  .3714286  3.9744511535877236    2.51210689875867  5.436795408416778
1410 110  .3492064  3.9437167422282013   2.473638267386958 5.4137952170694446
1440 100  .3174603  4.5122724855936855   2.935768996508071    6.0887759746793
1470  96  .3047619   4.679545651356969   2.915008202769464  6.444083099944475
1500  96  .3047619   4.679545651356969   2.915008202769464  6.444083099944475
1530  93  .2952381   4.791127198504803  2.9017299378344097  6.680524459175197
1560  90  .2857143   4.535125545765046    2.68041710129408  6.389833990236013
1590  87  .2761905    4.56434819372004  2.7865120368652256  6.342184350574854
1620  85 .26984128   4.711305948546441   3.112012418437499  6.310599478655383
1650  83 .26349208   4.696330087243326  3.1017411347269785 6.2909190397596735
1680  82 .26031747  4.5760002676350275  2.9981085791176625  6.153891956152393
1710  80 .25396827   4.298243982053482  2.6962880990034757  5.900199865103489
1740  73 .23174603   4.386591953033458  2.5268607287775184  6.246323177289397
end

label variable estimate "Estimated coefficient"
label variable min95 " Confidence interval "
label variable max95 " 95% "
twoway (scatter estimate days,legend(region(col(white))) yscale(titlegap(2)) xscale(range(0 1800) titlegap(2)) msymbol(square) msize(small)) (rcap min95 max95 days, lcolor(blue)), xlabel(0(250)1750, valuelabels angle(0) labsize(small)) xtitle("Number of days") ytitle("Grade variation (%)") ysca(titlegap(2)) ylab(, labs(small)) xsca(titlegap(2)) yline(0, lcolor(red)) graphregion(color(white)) ylabel(-2 (2) 8)

When I try to add the number of treated as a label, using mlabel(treated), it is sloppy:

label variable estimate "Estimated coefficient"
label variable min95 " Confidence interval "
label variable max95 " 95% "
twoway (scatter estimate days, mlabel(treated) legend(region(col(white))) yscale(titlegap(2)) xscale(range(0 1800) titlegap(2)) msymbol(square) msize(small)) (rcap min95 max95 days, lcolor(blue)), xlabel(0(250)1750, valuelabels angle(0) labsize(small)) xtitle("Number of days") ytitle("Grade variation (%)") ysca(titlegap(2)) ylab(, labs(small)) xsca(titlegap(2)) yline(0, lcolor(red)) graphregion(color(white)) ylabel(-2 (2) 8)

Any suggestions making this more aesthetic? I wanted to overlay a density plot using variable “proportion”, but I always fail. Could someone help me, please?