Overlapping plots for 2 periods

[Anouk Hoek]

Hello, I am having difficulties with a plot I am trying to make. I have data for 2 periods, for which I created a variable:

gen period = .
replace period = 1 if (dofc(datetime) >= td(1aug2022) & dofc(datetime) <= td(31mar2023))
replace period = 2 if (dofc(datetime) >= td(1aug2023) & dofc(datetime) <= td(31mar2024))

I then collapsed the data to get daily means for my variables:

* Create a variable for the date (without time)
gen date = dofc(datetime)
format %td date
* Calculate daily averages for each pollutant
collapse (mean) pollutant, by(date location_name period)

* Example generated by -dataex-. For more info, type help dataex
clear
input str29 location_name float(period date) double pollutant
"AmsterdamHaarlemmerwegNOx"     1 22858 25.408249999999995
"AmsterdamJanvanGalenstraatNOx" 1 22858 19.104416666666662
"AmsterdamStadhouderskadeNOx"   1 22858  26.94279166666667
"AmsterdamVanDiemenstraatNOx"   1 22858 18.140708333333336
"AmsterdamHaarlemmerwegNOx"     1 22859  51.79487499999999
"AmsterdamJanvanGalenstraatNOx" 1 22859             34.751
"AmsterdamStadhouderskadeNOx"   1 22859          30.427875
"AmsterdamVanDiemenstraatNOx"   1 22859          29.260125
"AmsterdamHaarlemmerwegNOx"     1 22860  62.93816666666667
"AmsterdamJanvanGalenstraatNOx" 1 22860  43.93387499999999
"AmsterdamStadhouderskadeNOx"   1 22860 34.135625000000005
"AmsterdamVanDiemenstraatNOx"   1 22860  37.19404166666667
"AmsterdamHaarlemmerwegNOx"     1 22861  48.97095833333333
"AmsterdamJanvanGalenstraatNOx" 1 22861           40.59175
"AmsterdamStadhouderskadeNOx"   1 22861 40.965833333333336
"AmsterdamVanDiemenstraatNOx"   1 22861  31.92554166666666
"AmsterdamHaarlemmerwegNOx"     1 22862           24.55225
"AmsterdamJanvanGalenstraatNOx" 1 22862 13.993291666666664
"AmsterdamStadhouderskadeNOx"   1 22862  21.94716666666667
"AmsterdamVanDiemenstraatNOx"   1 22862 15.096708333333334
"AmsterdamHaarlemmerwegNOx"     1 22863  52.17295833333333
"AmsterdamJanvanGalenstraatNOx" 1 22863            27.3935
"AmsterdamStadhouderskadeNOx"   1 22863            38.8675
"AmsterdamVanDiemenstraatNOx"   1 22863  20.07716666666667
"AmsterdamHaarlemmerwegNOx"     1 22864            61.8705
"AmsterdamJanvanGalenstraatNOx" 1 22864  33.35783333333334
"AmsterdamStadhouderskadeNOx"   1 22864  34.70204166666667
"AmsterdamVanDiemenstraatNOx"   1 22864 18.903249999999996
"AmsterdamHaarlemmerwegNOx"     1 22865 31.160708333333332
"AmsterdamJanvanGalenstraatNOx" 1 22865  28.55679166666667
"AmsterdamStadhouderskadeNOx"   1 22865  36.57366666666667
"AmsterdamVanDiemenstraatNOx"   1 22865  15.43833333333333
"AmsterdamHaarlemmerwegNOx"     1 22866 19.824166666666667
"AmsterdamJanvanGalenstraatNOx" 1 22866          15.191125
"AmsterdamStadhouderskadeNOx"   1 22866  38.65404166666666
"AmsterdamVanDiemenstraatNOx"   1 22866 11.907083333333333
"AmsterdamHaarlemmerwegNOx"     1 22867 14.210708333333336
"AmsterdamJanvanGalenstraatNOx" 1 22867 13.104083333333335
"AmsterdamStadhouderskadeNOx"   1 22867 28.665166666666664
"AmsterdamVanDiemenstraatNOx"   1 22867 12.432708333333336
"AmsterdamHaarlemmerwegNOx"     1 22868 20.532041666666668
"AmsterdamJanvanGalenstraatNOx" 1 22868 17.529125000000004
"AmsterdamStadhouderskadeNOx"   1 22868           31.42675
"AmsterdamVanDiemenstraatNOx"   1 22868          17.015125
"AmsterdamHaarlemmerwegNOx"     1 22869 22.865958333333335
"AmsterdamJanvanGalenstraatNOx" 1 22869 18.561833333333336
"AmsterdamStadhouderskadeNOx"   1 22869  32.60179166666667
"AmsterdamVanDiemenstraatNOx"   1 22869 18.604916666666664
"AmsterdamHaarlemmerwegNOx"     1 22870  70.42529166666665
"AmsterdamJanvanGalenstraatNOx" 1 22870  40.00958333333334
"AmsterdamStadhouderskadeNOx"   1 22870 31.134041666666672
"AmsterdamVanDiemenstraatNOx"   1 22870  34.72895833333333
"AmsterdamHaarlemmerwegNOx"     1 22871 30.570041666666658
"AmsterdamJanvanGalenstraatNOx" 1 22871 19.245500000000003
"AmsterdamStadhouderskadeNOx"   1 22871  20.79216666666667
"AmsterdamVanDiemenstraatNOx"   1 22871           22.36775
"AmsterdamHaarlemmerwegNOx"     1 22872  57.70708333333334
"AmsterdamJanvanGalenstraatNOx" 1 22872 44.643499999999996
"AmsterdamStadhouderskadeNOx"   1 22872 33.059583333333336
"AmsterdamVanDiemenstraatNOx"   1 22872  33.79945833333333
"AmsterdamHaarlemmerwegNOx"     1 22873  50.32695833333333
"AmsterdamJanvanGalenstraatNOx" 1 22873  38.40058333333334
"AmsterdamStadhouderskadeNOx"   1 22873  34.20541666666667
"AmsterdamVanDiemenstraatNOx"   1 22873             29.693
"AmsterdamHaarlemmerwegNOx"     1 22874          33.363875
"AmsterdamJanvanGalenstraatNOx" 1 22874 26.531125000000003
"AmsterdamStadhouderskadeNOx"   1 22874 30.497208333333333
"AmsterdamVanDiemenstraatNOx"   1 22874 18.123083333333337
"AmsterdamHaarlemmerwegNOx"     1 22875  41.52958333333333
"AmsterdamJanvanGalenstraatNOx" 1 22875  27.20820833333333
"AmsterdamStadhouderskadeNOx"   1 22875 30.805333333333333
"AmsterdamVanDiemenstraatNOx"   1 22875 21.364999999999995
"AmsterdamHaarlemmerwegNOx"     1 22876  58.07566666666667
"AmsterdamJanvanGalenstraatNOx" 1 22876 41.393375000000006
"AmsterdamStadhouderskadeNOx"   1 22876  33.58908333333333
"AmsterdamVanDiemenstraatNOx"   1 22876  32.84766666666667
"AmsterdamHaarlemmerwegNOx"     1 22877  52.10570833333332
"AmsterdamJanvanGalenstraatNOx" 1 22877  32.31462500000001
"AmsterdamStadhouderskadeNOx"   1 22877  27.93466666666667
"AmsterdamVanDiemenstraatNOx"   1 22877 23.502208333333332
"AmsterdamHaarlemmerwegNOx"     1 22878  42.36108333333334
"AmsterdamJanvanGalenstraatNOx" 1 22878 25.066375000000008
"AmsterdamStadhouderskadeNOx"   1 22878 31.159416666666672
"AmsterdamVanDiemenstraatNOx"   1 22878 24.291958333333337
"AmsterdamHaarlemmerwegNOx"     1 22879  39.05333333333333
"AmsterdamJanvanGalenstraatNOx" 1 22879 31.203999999999997
"AmsterdamStadhouderskadeNOx"   1 22879 19.156583333333334
"AmsterdamVanDiemenstraatNOx"   1 22879           32.60225
"AmsterdamHaarlemmerwegNOx"     1 22880           63.09825
"AmsterdamJanvanGalenstraatNOx" 1 22880 52.655791666666666
"AmsterdamStadhouderskadeNOx"   1 22880  33.80950000000001
"AmsterdamVanDiemenstraatNOx"   1 22880  37.98970833333333
"AmsterdamHaarlemmerwegNOx"     1 22881  53.02283333333332
"AmsterdamJanvanGalenstraatNOx" 1 22881  40.19837500000001
"AmsterdamStadhouderskadeNOx"   1 22881 36.459083333333325
"AmsterdamVanDiemenstraatNOx"   1 22881           34.36575
"AmsterdamHaarlemmerwegNOx"     1 22882  53.15616666666668
"AmsterdamJanvanGalenstraatNOx" 1 22882  43.89016666666666
"AmsterdamStadhouderskadeNOx"   1 22882  30.31045833333333
"AmsterdamVanDiemenstraatNOx"   1 22882 53.116125000000004
end
format %td date

I now want to make a separate plot for each location where those 2 periods overlap. But I get 2 separate graphs next to each other using this code:

twoway (line pollutant date if period == 1 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(blue) lwidth(medium)) ///
       (line pollutant date if period == 2 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(red) lwidth(medium)), ///
       title("Daily Averages for AmsterdamStadhouderskadeNOx") ///
       legend(label(1 "Aug 2022 - Mar 2023") label(2 "Aug 2023 - Mar 2024")) ///
       ytitle("(mean) pollutant") xtitle("date") ///
       xlabel(, angle(45))

I tried a couple of things, but they all do not seem to work. Can you guys help me? Thanks in advance!

[Chen Samulsion]

Maybe you can specify axes by using xaxis option. However, I think it's a misleading solution.

twoway (line pollutant date if period == 1 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(blue) lwidth(medium) xaxis(1)) ///
       (line pollutant date if period == 2 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(red) lwidth(medium) xaxis(2) xtitle("", axis(2))), ///
       title("Daily Averages for AmsterdamStadhouderskadeNOx") ///
       legend(label(1 "Aug 2022 - Mar 2023") label(2 "Aug 2023 - Mar 2024")) ///
       ytitle("(mean) pollutant") xtitle("date")

[Nick Cox]

I don't understand the question. The periods don't overlap. 1 is [td(1aug2022), td(31mar2023)] and 2 is [td(1aug2023), td(31mar2024)].

Even if you did have overlapping periods, so that observations qualify for category 1 and category 2, the code would assign them to category 2/

[Anouk Hoek]

I don't understand the question. The periods don't overlap. 1 is [td(1aug2022), td(31mar2023)] and 2 is [td(1aug2023), td(31mar2024)].

Even if you did have overlapping periods, so that observations qualify for category 1 and category 2, the code would assign them to category 2/

Sorry I probably did not explain my question properly. The periods don't overlap indeed, but I want the graphs of the 2 periods to overlap, to see the difference in fluctuations for the 2 periods

[Nick Cox]

I see! Calculate a new time variable which is days since the previous 31 July.

[Anouk Hoek]

I see! Calculate a new time variable which is days since the previous 31 July.

Hi Nick, can you help me a bit with this? I don't see how I would use this in the plot..

[Nick Cox]

gen date2 = cond(period == 1, date - dmy(31,7,2022), cond(period == 2, date - dmy(31, 7, 2023), .))


twoway (line pollutant date2 if period == 1 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(blue) lwidth(medium)) ///
       (line pollutant date2 if period == 2 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(red) lwidth(medium)), ///
       title("Daily Averages for AmsterdamStadhouderskadeNOx") ///
       legend(label(1 "Aug 2022 - Mar 2023") label(2 "Aug 2023 - Mar 2024")) ///
       ytitle("(mean) pollutant") xtitle("days after 31 July") ///
       xlabel(1 "1 Aug" 32 "1 Sep" 62 "1 Oct" 93 "1 Nov" 123 "1 Dec" 154 "1 Jan" 185 "1 Feb" 213 "1 Mar")

The axis labels aren't exactly right given 29 February 2024, but perhaps close enough for graphical purposes.

[Anouk Hoek]

Maybe you can specify axes by using xaxis option. However, I think it's a misleading solution.

twoway (line pollutant date if period == 1 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(blue) lwidth(medium) xaxis(1)) ///
(line pollutant date if period == 2 & location_name == "AmsterdamStadhouderskadeNOx", lcolor(red) lwidth(medium) xaxis(2) xtitle("", axis(2))), ///
title("Daily Averages for AmsterdamStadhouderskadeNOx") ///
legend(label(1 "Aug 2022 - Mar 2023") label(2 "Aug 2023 - Mar 2024")) ///
ytitle("(mean) pollutant") xtitle("date")

I overlooked your suggestion, but this worked! Thank you so much

[Nick Cox]

My guess is with #2 perfect alignment is not guaranteed without extra work. With #7 it is automatic.