produces a plot that is identical to my third example (not second, as
I said).
There's no need to define local macros for red and green;
-mcolor(red)- and -mcolor(green)- work just fine.
But, as shown in my third example, when we use two separate -scatter-
statements the sizes of the plot symbols are computed separately for
each statement, so that the areas of the resulting symbols are no
longer comparable between the groups (foreign and domestic).
This is why I'd like instead to specify the color via a variable and
use a single -scatter- statement. This would be analogous to the
-mlabel(rep78)- syntax in the first example.
Thanks for any solution
Jake Wegelin
On Sun, Apr 26, 2009 at 7:20 PM, Jacob Wegelin <[email protected]> wrote:
> I believe this is the same as my second example. The two separate
> -scatter- statements cause the sizes of the plotting circles to be
> recalculated for each -scatter- statement, so that the areas are no
> longer proportional to the sample size N. This is why I want to
> specify the color via a variable, analogous to the -msymbol(rep78)-
> example.
>
> Jake
>
> On Sun, Apr 26, 2009 at 6:58 PM, Kieran McCaul <[email protected]> wrote:
>> .
>>
>>
>> How about:
>>
>> local f0 = "red"
>> local f1 = "green"
>> twoway (scatter mpg rep78 [fweight=N] if foreign==0, msymbol(Oh)
>> mcolor(`f0')) ///
>> (scatter mpg rep78 [fweight=N] if foreign==1, msymbol(Oh)
>> mcolor(`f1')) ///
>> , legend(off)
>>
>>
>>
>>
>> ______________________________________________
>> Kieran McCaul MPH PhD
>> WA Centre for Health & Ageing (M573)
>> University of Western Australia
>> Level 6, Ainslie House
>> 48 Murray St
>> Perth 6000
>> Phone: (08) 9224-2701
>> Fax: (08) 9224 8009
>> email: [email protected]
>> http://myprofile.cos.com/mccaul
>> http://www.researcherid.com/rid/B-8751-2008
>> ______________________________________________
>> Epidemiology is so beautiful and provides such an important perspective
>> on human life and death,
>> but an incredible amount of rubbish is published. Richard Peto (2007)
>>
>> -----Original Message-----
>> From: [email protected]
>> [mailto:[email protected]] On Behalf Of Jacob Wegelin
>> Sent: Monday, 27 April 2009 6:30 AM
>> To: [email protected]
>> Subject: st: -twoway scatter- different colors for different
>> observations
>>
>> The following syntax specifies a plot marker or -mlabel- that can take
>> on a different value at each plotting point (at each observation),
>> according to the value of variable specified in the -mlabel-
>> statement.
>>
>> clear all
>> sysuse auto
>> scatter price mpg, mlabel(rep78) m(i) mlabposition(3)
>>
>> But I would like to define a variable that specifies that certain
>> observations be plotted green, others red. Alternatively that certain
>> be plotted with color -none- and others with a visible color. How does
>> one do this?
>>
>> To motivate this (with an artificial example constructed to mimic a
>> real example): The following code plots average mpg by rep78, averaged
>> separately for foreign and domestic autos. Crucially, the area of the
>> plotting symbol is proportional to the sample size. I would like to
>> distinguish visually between domestic and foreign autos, though. Two
>> *separate* -scatter- statements (the second -twoway- command below)
>> don't give the desired result, because the plotting symbols are
>> re-scaled for each -scatter- statement. You can see this by switching
>> rapidly between the two exported graphs, junk1.pdf and junk2.pdf. And
>> an attempt to define a string variable -mycolor- which takes on values
>> "red" and "green", and then to specify -mcolor(mycolor)- analogously
>> to the -mlabel(rep78)- statement above, returns an error.
>>
>> clear all
>> set more on
>> sysuse auto
>> drop if rep78==.
>> sort foreign rep78
>> collapse (mean) mpg (count) N=price, by(foreign rep78)
>> /*
>> A larger value for N will make the problem easier to see.
>> */
>> replace N=50 in 6
>> list
>> set scheme lean1
>> twoway (scatter mpg rep78 [fweight=N], msymbol(Oh))
>> graph export junk1.pdf, replace
>> more
>> twoway ///
>> (scatter mpg rep78 [fweight=N] if foreign==0, msymbol(Oh)
>> mcolor(red)) ///
>> (scatter mpg rep78 [fweight=N] if foreign==1, msymbol(Oh)
>> mcolor(green)) ///
>> , legend(off)
>> graph export junk2.pdf, replace
>> more
>> /* The following returns an error */
>> gen mycolor=""
>> replace mycolor="red" if foreign==0
>> replace mycolor="green" if foreign==1
>> tabulate foreign mycolor
>> twoway (scatter mpg rep78 [fweight=N], msymbol(Oh) mcolor(mycolor))
>>
>>
>> The [G] GRAPHICS manual under -marker_options- says that one could
>> define color by specifying a list of elements, as
>>
>> -mcolor( red green red)-
>>
>> but this would be clumsy and error-prone. There must be a way to use
>> the values of a variable, as in the -mlabel(rep78)- example?
>>
>> Thanks for any insights
>>
>> Jacob A. Wegelin
>> Assistant Professor
>> Department of Biostatistics
>> Virginia Commonwealth University
>> 730 East Broad Street Room 3006
>> P. O. Box 980032
>> Richmond VA 23298-0032
>> U.S.A.
>> E-mail: [email protected]
>> URL: http://www.people.vcu.edu/~jwegelin
>> *
>> * 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/
>>
>>
>>
>> *
>> * 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/
>>
>
*
* 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/
