Re: scatter plot w/ colors

Hi Alessandra,

I see what you mean. Looking at the coding of GetFillColor, I didn't
realize that it wants a complete colormap to be input, with the
foreground and background colors included. That adds a bit of code to
what I sent you last night, and I tweaked a couple other lines after
testing out what I sent you yesterday.

Here's what I have for my test script:

pres = True
z = random_uniform(-50,50,(/100/)) ; random timeseries

arr = nice_mnmxintvl(min(z),max(z),21,False)
;--------round added to get rid of Argument type mismatch error
sorted_z = fspan(arr(0),arr(1),round(((arr(1)-arr(0))/arr(2))+1,3))

wks = gsn_open_wks("x11","test")
gsn_define_colormap(wks,"BlWhRe")
cmap = gsn_retrieve_colormap(wks)
ncolors = dimsizes(cmap(:,0)) ; number of colors

c_ind = round(fspan(2,ncolors-1,dimsizes(sorted_z)+1),3) ; round to int
ncmap = new((/dimsizes(c_ind)+2,3/),typeof(cmap)) ; add black + white
ncmap(:1,:) = cmap(:1,:)
ncmap(2:,:) = cmap(c_ind,:)
delete(cmap)

do j=0,dimsizes(z)-1
   pres_at_gsMarkerColor=GetFillColor(sorted_z,ncmap,z(j))
   print(pres_at_gsMarkerColor)
; gsn_polymarker(wks,plota,x,y,pres)
end do

Notice I'm not plotting anything, I'll leave that to you. Let me know if
this version doesn't work as advertised..
Adam

Alessandra Giannini wrote:
>
> Adam,
>
> thanks for your reply, and example. I can follow what you are doing, but
> I still get this complaint:
>
> "Not enough colors in colormap for number of contour levels"
>
> The line is repeated many times - too many for me to try and count
> them... ;-)
>
> I checked the dimensions of the vectors, and even when I reduce the z,
> sorted_z vectors to be smaller in dimension than the original color map
> it gives me the same error.
> The problem is in the line where the "GetFillColor" function is applied
> - if I comment that out it runs, but obviously does not produce the
> desired output.
>
> Do you have any insights?
>
> Thanks very much!
> best, alessandra
>
>
>
>
>
>
> On Jan 15, 2009, at 6:09 PM, Adam Phillips wrote:
>
>> Hi Alessandra,
>>
>> If you wanted to automatically come up with your contour (for lack of
>> a better word) levels, you can use nice_mnmxintvl first:
>> http://www.ncl.ucar.edu/Document/Functions/Built-in/nice_mnmxintvl.shtml
>>
>> arr = nice_mnmxintvl(min(z),max(z),21,False)
>> sorted_z = fspan(arr(0),arr(1),((arr(1)-arr(0))/arr(2))+1)
>>
>> Now let's say you wanted to use the BlWhRe colormap:
>> http://www.ncl.ucar.edu/Document/Graphics/ColorTables/BlWhRe.shtml
>>
>> wks_define_colormap(wks,"BlWhRe")
>> cmap = gsn_retrieve_colormap(wks)
>> dimz = dimsizes(cmap)
>> ncolors = dimz(0) ; number of colors
>> c_ind = round(fspan(2,ncolors-1,dimsizes(sorted_z)+1),3) ; round to int
>> do j=0,dimsizes(z)-1
>> pres_at_gsMarkerColor=GetFillColor(sorted_z,cmap(c_ind,:),z(j))
>> gsn_polymarker(wks,plota,x,y,pres)
>> end do
>>
>> (untested) I think that will do it.. If that doesn't do what you want,
>> let us know..
>> Adam
>>
>> Alessandra Giannini wrote:
>>> Hi everyone,
>>> I am trying to figure out how to plot a scatterplot of THREE
>>> variables in the following way:
>>> the first variable runs along the x-axis, the second along the y-axis
>>> and the third adds variation in color to the points.
>>> To give you a prototypical example, here is a plot of average
>>> temperature/precipitation colored by elevation in stations around the
>>> central Mediterranean, computed in the IRI Data Library:
>>> <http://iridl.ldeo.columbia.edu/expert/SOURCES/.NOAA/.NCDC/.GCPS/.MONTHLY/.STATION/lon/%280%29%2820%29masknotrange/SELECT/lat/%2835%29%2845%29masknotrange/SELECT/mean/temp%5BT%5Daverage/1/index/.prcp%5BT%5Daverage/3/index/.elev/fig://XOVY/1/def/scattercolor/:fig/>
>>>
>>> In NCL I tried this way a while back - start with the scatterplot of
>>> 2 variables:
>>> plot = gsn_csm_xy(wks,x,y,res)
>>> and then add color:
>>> pres = True
>>> pres_at_gsMarkerIndex = (/ 16 /)
>>> pres_at_gsMarkerSizeF = (/ 0.01 /)
>>> do j=0,dimsizes(z)-1
>>> pres_at_gsMarkerColor=GetFillColor(sorted_z,cmap,z(j))
>>> gsn_polymarker(wks,plota,x,y,pres) end do
>>> where cmap is defined from an array or RGB values, stored in an
>>> external file:
>>> cmap = RGBtoCmap("csimple.txt")
>>> that typically looks like this:
>>> 255 255 255
>>> 0 0 0
>>> 0 0 128
>>> 0 0 205
>>> 30 144 255
>>> 0 191 255
>>> 127 255 212
>>> 255 255 255
>>> 255 255 0
>>> 255 215 0
>>> ...
>>> My problem is in defining a color scale automatically, in the sense
>>> that I would like to be able to just pick out a color map regardless
>>> of the number of points in the scatter - one of the classic color
>>> scales defined in ncl would do, like BlueRed.
>>> Can I do that? And if so, how?
>>> Your help is much appreciated.
>>> thanks, alessandra
>>> --
>>> Alessandra Giannini
>>> IRI for Climate and Society - The Earth Institute at Columbia University
>>> P.O. Box 1000, Palisades NY 10964-8000
>>> phone/fax: +1 845 680-4473/4864 - email: alesall_at_iri.columbia.edu
>>> <mailto:alesall_at_iri.columbia.edu>
>>> ------------------------------------------------------------------------
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>> --
>> --------------------------------------------------------------
>> Adam Phillips asphilli_at_ucar.edu
>> National Center for Atmospheric Research tel: (303) 497-1726
>> ESSL/CGD/CAS fax: (303) 497-1333
>> P.O. Box 3000
>> Boulder, CO 80307-3000 http://www.cgd.ucar.edu/cas/asphilli

-- 
--------------------------------------------------------------
Adam Phillips			             asphilli_at_ucar.edu
National Center for Atmospheric Research   tel: (303) 497-1726
ESSL/CGD/CAS                               fax: (303) 497-1333
P.O. Box 3000				
Boulder, CO 80307-3000	  http://www.cgd.ucar.edu/cas/asphilli
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk