Two y-axes plot

[Budu Gulo]

My data set has three variables: date, type_of_company, and sales. The variable type_of_company is a categorical variable, with four categories : "A+", "A-", "B+", and "B". I want to generate two plots in one graph: One plot would present sales of "A+" companies by month; other one would present percentage of sales by "A+" companies by month (that is, if we add up all the sales by month what percent belongs to "A+" companies).

In short, the x-axis would represent months, and the two y-axes --left and right--would represent sales and percent of sales of "A+" companies.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float date str2 type_of_company float sales
17167 "A-"  461390
17198 "A-"  484773
17226 "A-"  475477
17257 "A-"  487706
17287 "A-"  508125
17318 "A-"  530895
17348 "A-"  528911
17379 "A-"  532296
17410 "A-"  528430
17440 "A-"  577063
17471 "A-"  569910
17501 "A-"  556396
17532 "A-"  554090
17563 "A-"  527268
17592 "A-"  516887
17623 "A-"  512096
17653 "A-"  529073
17684 "A-"  515582
17714 "A-"  451889
17745 "A-"  454076
17776 "A-"  473585
17806 "A-"  403169
17837 "A-"  336504
17867 "A-"  282454
17167 "B+"   41765
17198 "B+"   44656
17226 "B+"   44798
17257 "B+"   49307
17287 "B+"   54469
17318 "B+"   57547
17348 "B+"   64375
17379 "B+"   64850
17410 "B+"   65161
17440 "B+"   74496
17471 "B+"   87354
17501 "B+"   84085
17532 "B+"   88423
17563 "B+"   72580
17592 "B+"   69048
17623 "B+"   74501
17653 "B+"   80657
17684 "B+"   87685
17714 "B+"   78898
17745 "B+"  108294
17776 "B+"  108653
17806 "B+"   81871
17837 "B+"   57808
17867 "B+"   44567
17167 "B"  1042719
17198 "B"  1060750
17226 "B"  1051083
17257 "B"  1062208
17287 "B"  1076744
17318 "B"  1113599
17348 "B"  1163229
17379 "B"  1125650
17410 "B"  1103981
17440 "B"  1174523
17471 "B"  1250127
17501 "B"  1152243
17532 "B"  1116859
17563 "B"  1039562
17592 "B"  1016965
17623 "B"   995591
17653 "B"  1018994
17684 "B"  1014310
17714 "B"   901925
17745 "B"   866273
17776 "B"   862769
17806 "B"   791357
17837 "B"   667862
17867 "B"   537538
17167 "A+"  689146
17198 "A+"  703882
17226 "A+"  685212
17257 "A+"  710810
17287 "A+"  740271
17318 "A+"  771838
17348 "A+"  764078
17379 "A+"  752530
17410 "A+"  767773
17440 "A+"  811376
17471 "A+"  841409
17501 "A+"  830936
17532 "A+"  822536
17563 "A+"  775964
17592 "A+"  789393
17623 "A+"  794566
17653 "A+"  758239
17684 "A+"  753736
17714 "A+"  699446
17745 "A+"  697625
17776 "A+"  703192
17806 "A+"  674108
17837 "A+"  556932
17867 "A+"  473251
end
format %td date

[Carole J. Wilson]

You don't say what kind of graph you want, so I'm assuming a line graph. I am also assuming you only have only one "A+" company per month. If you have multiples of the same type, then the code will need to be altered.

gen monthly=mofd(date)
format %tm monthly
bysort monthly: egen all_sales=total(sales)
gen pc_sales=(sales/all_sales)*100

sort monthly
twoway line sales monthly if type_of_company=="A+" || line pc_sales monthly if type_of_company=="A+" , yaxis(2)

[Budu Gulo]

I was indeed looking for a line graph. Thanks a lot Carole J. Wilson!