Re: st: additional lines in a plot

[Nick Cox <njcoxstata@gmail.com>]

According to your example, -Austria- is a string variable. -summarize-
is not illegal with a string variable but attempts to use r(min) and
r(max) thereafter will be rewarded with missing values. Remarkably, or
not, this doesn't bite you until your graph command. In terms of what
you were doing before, you should use -dAustria-.

Nick

On Sat, Jun 9, 2012 at 3:24 PM, stef salvez <loggyedy@googlemail.com> wrote:
> Thank you for this Nick,
>
> I add to my initial code  what you proposed; that is,
>
> su Austria, meanonly
>
> local max = r(max)
> local date = r(min)
> local dates `date'
>
> while `date' < `max' {
>         local date = `date' + 28
>         local dates `dates' `date'
> }
>
> and works perfectly.
>
>
> I only get a warning sign when I then type
>
>
> scatter Aus  dAustria,   xline(`dates', lstyle(grid)  lw(vvthin) lcolor(gs12))
>
> invalid line argument, .
> r(198);
>
> Do I miss something?
>
> On 6/9/12, Nick Cox <njcoxstata@gmail.com> wrote:
>> Various ways depending on your rule for which you want plotted.
>>
>> But here is one.
>>
>> su datevar, meanonly
>>
>> local max = r(max)
>> local date = r(min)
>> local dates `date'
>>
>> while `date' < `max' {
>>           local date = `date' + 28
>>           local dates `dates' `date'
>> }
>>
>> ... xli(`dates', lstyle(grid)  lw(vvthin) lcolor(gs12))
>>
>> But judging from your example data this will just make a busy graph
>> busier. I would think in terms of plotting
>>
>> (time since last date) vs (this date)
>>
>> to make the gaps more evident.
>>
>> Nick
>>
>> On Fri, Jun 8, 2012 at 10:29 PM, stef salvez <loggyedy@googlemail.com>
>> wrote:
>>> But I do not know how to add these vertical lines every 28 dates. Some
>>> help would be greatly appreciated. It is an advanced topic and you are
>>> the right person to provide some information
>>>
>>> On 6/8/12, stef salvez <loggyedy@googlemail.com> wrote:
>>>> Thanks Nick.
>>>>
>>>> As you said I added xline() for one time point for Austria:
>>>>
>>>>
>>>>   scatter Aus  dAustria, xsize(20) ysize(8) c(l l)  xline( 17922 ) ||///
>>>>  scatter Bel dBelgium , c(l l)    || ///
>>>>  scatter Aus t10, ms(none) mla(dA2) mlabpos(910) || ///
>>>>  scatter Bel t9 , ms(none) mla(dB2) mlabpos(910) ///
>>>>  yla(9 "Belgium"  10 "Austria"    , ang(h)) ///
>>>>  xla(`=mdy(10,5,2008)' `=mdy(1,1,2010)' `=mdy(1,1,2012)', format(%td))
>>>> legend(off)
>>>>
>>>> and it worked. The only difficulty is that I have to insert manually
>>>> all these numbers (like 17922) which are the numerical representation
>>>> of the dates. IS there any quicker way to do that; as I have many
>>>> countries and for each country I have a different a sequence of dates.
>>>> One example would be sufficient for me to understand how to move on,
>>>>
>>>>
>>>> thanks again
>>>>
>>>>
>>>> On 6/8/12, Nick Cox <njcoxstata@gmail.com> wrote:
>>>>> -xaxis()- is an option, not a command, but of no help here. You want
>>>>> -xline()-. See also the SJ Tip cited earlier today to you.
>>>>>
>>>>> Nick
>>>>>
>>>>> On Fri, Jun 8, 2012 at 6:30 PM, stef salvez <loggyedy@googlemail.com>
>>>>> wrote:
>>>>>
>>>>>>  According to the dataset below i have observations every 28 days (4
>>>>>> weeks)  but in some cases i have jumps ( 35 days-or 5 weeks- and 42
>>>>>> days -or 6 weeks).
>>>>>>
>>>>>>  I would like to construct vertical gridlines   every 28 days so as
>>>>>> to see easier the one- and/or two-week shifts. The solution mignt be
>>>>>> the
>>>>>>  -xaxis- command but i do not know exactly how to fit it to the code
>>>>>> below
>>>>>>
>>>>>>
>>>>>> clear all
>>>>>> cd d:\
>>>>>> input str8  (Austria           Belgium          )
>>>>>>  "2/11/08"     "07/12/08"
>>>>>> "30/11/08"          "04/01/09"
>>>>>> "28/12/08"          "01/02/09"
>>>>>> "25/01/09"          "01/03/09"
>>>>>> "22/02/09"          "29/03/09"
>>>>>> "22/03/09"          "26/04/09"
>>>>>> "19/04/09"          "24/05/09"
>>>>>> "17/05/09"          "21/06/09"
>>>>>> "14/06/09"          "19/07/09"
>>>>>> "12/07/09"          "16/08/09"
>>>>>> "09/08/09"          "13/09/09"
>>>>>> "06/09/09"          "11/10/09"
>>>>>> "04/10/09"          "08/11/09"
>>>>>> "01/11/09"          "06/12/09"
>>>>>> "29/11/09"          "03/01/10"
>>>>>> "27/12/09"          "31/01/10"
>>>>>> "31/01/10"          "28/02/10"
>>>>>> "28/02/10"          "28/03/10"
>>>>>> "28/03/10"          "25/04/10"
>>>>>> "25/04/10"          "23/05/10"
>>>>>> "23/05/10"          "20/06/10"
>>>>>> "20/06/10"          "18/07/10"
>>>>>> "18/07/10"          "15/08/10"
>>>>>> "15/08/10"          "12/09/10"
>>>>>> "12/09/10"          "10/10/10"
>>>>>> "10/10/10"          "07/11/10"
>>>>>> "07/11/10"          "05/12/10"
>>>>>> "05/12/10"          "02/01/11"
>>>>>> "01/01/11"          "30/01/11"
>>>>>> "30/01/11"          "27/02/11"
>>>>>> "27/02/11"          "27/03/11"
>>>>>> "27/03/11"          "24/04/11"
>>>>>> "24/04/11"          "22/05/11"
>>>>>> "22/05/11"          "19/06/11"
>>>>>> "19/06/11"          "17/07/11"
>>>>>> "17/07/11"          "14/08/11"
>>>>>> "14/08/11"          "11/09/11"
>>>>>> "11/09/11"          "09/10/11"
>>>>>> "09/10/11"          "06/11/11"
>>>>>>
>>>>>> end
>>>>>> gen dAustria = date(Austria, "DM20Y")
>>>>>> gen dBelgium = date(Belgium, "DM20Y")
>>>>>>
>>>>>> gen dA2 = dAustria - dAustria[_n-1]
>>>>>> gen dB2 = dBelgium - dBelgium[_n-1]
>>>>>>  gen Aus= 10
>>>>>> gen Bel = 9
>>>>>> gen t10 = (dAustria + dAustria[_n-1]) / 2
>>>>>> gen t9 = (dBelgium + dBelgium[_n-1]) / 2
>>>>>>
>>>>>>  scatter Aus  dAustria, xsize(20) ysize(8) c(l l)|| scatter Bel
>>>>>> dBelgium , c(l l)    || //
>>>>>> scatter Aus t10, ms(none) mla(dA2) mlabpos(910) || ///
>>>>>> scatter Bel t9 , ms(none) mla(dB2) mlabpos(910)
>>>>>>  yla(9 "Belgium"  10 "Austria"    , ang(h)) xla(, format(%td))
>>>>>> legend(off)

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/