Re: st: Heatmaps

[Keith Dear <keithdear4@gmail.com>]

Oops.
Apologies for another inadvertence, sending an attachment to the list.
Hope it brightened everyone's day! It's risky having two gmail windows
open at once, I find -- drafts get conflicted.
kd

On 17 March 2011 10:29, Keith Dear <keithdear4@gmail.com> wrote:
> Hi Austin,
> Referring to tddens3 was inadvertent in my post -- I merely added an
> option SAVING(string) to write out the grid of densities as a .csv
> file -- useful for plotting in other programs (I'll send it privately
> to brighten your day).
>
> Other features? -- it would be nice to have "pretty" axis labels
> instead of the quartiles.
> Keith
>
>
>
> On 17 March 2011 03:32, Austin Nichols <austinnichols@gmail.com> wrote:
>> Keith Dear <keithdear4@gmail.com>:
>> I am joining this discussion late, but you are right that -tddens-
>> uses Square marker symbols that can overplot or not cover the plot
>> area completely.  I will investigate using -tw rbar- or -tw rarea- to
>> instead graph areas instead that adjoin perfectly.  Anyone who needs
>> this functionality can email me directly to nag me.  Keith--did you
>> modify tddens and save it as -tddens3- or did someone else give you a
>> modified copy?  What other features do you need from it?
>>
>> On Tue, Mar 8, 2011 at 2:49 AM, Keith Dear <keithdear4@gmail.com> wrote:
>>> They don't.
>>> -tddens- has an option Gridpoints that defaults to 40, and playing
>>> with it shows that the squares roughly abut for g=30, not 40.
>>> g<30 leaves gaps (and demonstrates a famous optical illusion).
>>> It's interesting that it still seems to work well with g>30, even
>>> though the squares must overlap, which I suppose distorts the image.
>>>
>>> clear all
>>> set obs 1000
>>> g x=rnormal(0,1)
>>> g y=(x+rnormal(0,1))/sqrt(2)
>>> drop if abs(x)<1 & abs(y)<1
>>> tddens3 y x, g(24) bw(.01)
>>>
>>> Keith
>>>
>>>
>>> On 7 March 2011 10:56, Brendan Halpin <brendan.halpin@ul.ie> wrote:
>>>> On Sun, Mar 06 2011, Keith Dear wrote:
>>>>
>>>>> Take a look inside Austin Nichols's very clever -tddens- (-findit- finds it).
>>>>
>>>> Almost unreadable at first glance but indeed very clever. I see the main
>>>> idea (round your Z variable, and overlay scatterplots of X and Y for
>>>> each level of rounded Z, with direct control of the colour via the
>>>> mcolor option -- that will allow Stata to do anything,
>>>> colour-palette-wise, that gnuplot can).
>>>>
>>>> What I can't see is how the code makes the square scatterplot symbols
>>>> abut exactly.
>>>>
>>>> Brendan
>>>> --
>>>> Brendan Halpin,  Department of Sociology,  University of Limerick,  Ireland
>>>> Tel: w +353-61-213147 f +353-61-202569 h +353-61-338562; Room F1-009 x 3147
>>>> mailto:brendan.halpin@ul.ie  http://www.ul.ie/sociology/brendan.halpin.html
>>
>> *
>> *   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/
>>
>
>
>
> --
> Dr Keith Dear
> National Centre for Epidemiology and Population Health
> ANU College of Medicine, Biology and Environment
> Australian National University
> Canberra, ACT 0200 Australia
> CRICOS provider #00120C
> Phone +61 (02) 6273 2208
> Mobile 0424 450 396
>



-- 
Dr Keith Dear
National Centre for Epidemiology and Population Health
ANU College of Medicine, Biology and Environment
Australian National University
Canberra, ACT 0200 Australia
CRICOS provider #00120C
Phone +61 (02) 6273 2208
Mobile 0424 450 396

*
*   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/