Dear Richard,
you have a nice collection of references of why it is neither
desirable, nor possible to do pattern fills in Stata. But if you still
think you need it, then see the example -parea- command:
findit parea
It does pattern fills (up to 10) in standard twoway area command. You
can modify the bar command in a similar manner. This works in Windows.
Not sure about other systems, but macs are almost surely out of play.
Alternatively, you can look in the settings of your printer to see if
it has any option to adjust half-toning used to represent shades (if
this is a bw printer, like a laserjet). I think you might have better
luck with the older drivers (but the patters if any will probably be
too small to be distinguishable with a naked eye, e.g. hatching at
1200 dpi....)
Hope this helps.
Best regards,
Sergiy Radyakin
On 7/17/08, Richard Zijdeman <[email protected]> wrote:
> Dear all,
>
> I had a question on how to provide patterns on bars in bar charts, in
> order to differentiate between different categories, when bars are
> printed in grayscale. This issue has been addressed on the statalist
> (see below), and it is recommended that you should look for another
> type of graph. For those who will also need to deal with this issue, I
> would like to add the following to that discussion:
>
> (1) The command "tabhbar" actually does provide such a graph with
> different patterned bars in color. (You can obtain it by typing: "ssc
> install tabhbar" on the command line.) But you will find it still
> difficult to interpret in grayscale, with more than 4 categories.
>
> (2) I am really convinced now, that you should indeed use another type
> of graph, and that "tabplot" is actually what you might be looking
> for. ("ssc install tabplot").
>
> I have a number of occupations, that are categorized in 7 groups:
> occ7. I also have three class schemes (class1-class3) in which all of
> these occupations are categorized according to "social economic
> class". I wanted to show, to what the degree the 7 occupational groups
> are categorized similarily over the three class schemes. Here's how to
> do it:
>
> tabplot occ7 class1, name(g_class1, replace)
> tabplot occ7 class2, name(g_class2, replace)
> tabplot occ7 class3, name(g_class3, replace)
> graph combine g_class1 g_class2 g_class3, row(1) xcommon
> * the row(1) makes sure that the graphs are put in 1 row, while
> xcommon makes sure that the classes have the same xaxis.
>
> For more info on tabplot and similar commands dealing with categorizations:
> http://www.stata-journal.com/sjpdf.html?articlenum=gr0004
>
> Thank you for tabplot and for the earlier responses on the statalist
> on this matter.
>
> Cheers,
>
> Richard
>
> -- Earlier postings on the statalist addressing this issue.
> *st: Fill patterns in stata8 hbar graph, Eric Magar (Fri Sep 23 17:00:14 2005)
> **Re: st: Fill patterns in stata8 hbar graph, Eric G. Wruck (Fri Sep
> 23 17:40:29 2005)
> ***Re: st: Fill patterns in stata8 hbar graph, Michael Blasnik (Fri
> Sep 23 18:40:04 2005)
> *RE: st: Fill patterns in stata8 hbar graph, Nick Cox (Sun Sep 25 10:40:05 2005)
> *st: RE: Making bar charts readable in a grayscale photocopy (Fri, 24
> Oct 2003 16:59:13 -0400)
> *st: Bar graph (Wed, 29 Oct 2003 12:17:21 -0500)
>
> --
> Richard L. Zijdeman
> Ph.D. Student
> ICS / Utrecht University
> Phone: +31 30 253 1967
> Fax: +31 30 253 4405
> Mobile: +31 6 2903 1953
> Skype: rlzijdeman
> *
> * 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/
