R help - Nov 2003 - barchart in lattice

Dear all,

I have two factors 'country' and 'status' which I would like to
plot via barchart (lattice).
'status' consist of three different levels and should be the grouping
variable, i.e. there
should be drawn three different panels and within each panel a barchart of
'country'.

barchart(daten$COUNTRY|daten$STATUS), 
barchart(table(daten$COUNTRY)|table(daten$STATUS)),
etc. are obviously wrong.  

I have absolutely no idea how to do this. 

You can download the datafile from
<http://www.uni-koeln.de/~ahf34/meta.txt>

TIA,

Bernd 

-- 
Bernd Weiss, M.A.
Universitaet zu Koeln / University of Cologne
Forschungsinstitut fuer Soziologie / Research Institute for Sociology
Greinstr. 2 / 50 939 Cologne / Germany
Phone: +49 221 / 470-4234
E-Mail: <bernd.weiss at uni-koeln.de>

On Sunday 02 November 2003 04:17, Bernd Weiss wrote:
> Dear all,
>
> I have two factors 'country' and 'status' which I would
like to plot via
> barchart (lattice). 'status' consist of three different levels and
should
> be the grouping variable, 
the correct terminology would be 'conditioning' variable. Grouping
variables
distinguish data within a panel.
> i.e. there should be drawn three different panels
> and within each panel a barchart of 'country'.
I'm not sure what that means. If you mean barchart of the frequency of 
occurrence of each country within each 'status', then you could try 

foo <- read.table("meta.txt")
barchart(daten.COUNTRY ~ Freq| daten.STATUS, 
         data = as.data.frame(table(foo)), 
         origin = 0)

HTH,
Deepayan

> barchart(daten$COUNTRY|daten$STATUS),
> barchart(table(daten$COUNTRY)|table(daten$STATUS)),
> etc. are obviously wrong.
>
> I have absolutely no idea how to do this.
>
> You can download the datafile from
> <http://www.uni-koeln.de/~ahf34/meta.txt>
>
> TIA,
>
> Bernd

Am 2 Nov 2003 um 10:34 hat Deepayan Sarkar geschrieben:
> 
> On Sunday 02 November 2003 04:17, Bernd Weiss wrote:
> > Dear all,
> >
> > I have two factors 'country' and 'status' which I
would like to plot via
> > barchart (lattice). 'status' consist of three different levels
and should
> > be the grouping variable, 
> 
> the correct terminology would be 'conditioning' variable. Grouping
variables
> distinguish data within a panel.
> 
> > i.e. there should be drawn three different panels
> > and within each panel a barchart of 'country'.
> 
> I'm not sure what that means. If you mean barchart of the frequency of 
> occurrence of each country within each 'status', then you could try
> 
> foo <- read.table("meta.txt")
> barchart(daten.COUNTRY ~ Freq| daten.STATUS, 
>          data = as.data.frame(table(foo)), 
>          origin = 0)
> 
That's it! Thank you very much for your help!

Bernd

Bernd,

for counts use histogram instead of barchart, and leave the left side of the
formula empty.

library(lattice)
meta<-read.table("meta.txt",header=T)
histogram(~daten.COUNTRY|daten.STATUS,data=meta)

Check "scales" in xyplot (it's also valid for histogram) so you
get the
names of the
countries as vertical x-labels instead of the numbers.

--------- You wrote ----
I have two factors 'country' and 'status' which I would like to
plot via
barchart (lattice). 'status' consist of three different levels and
should be
the grouping variable, i.e. there should be drawn three different panels and
within each panel a barchart of 'country'.

barchart(daten$COUNTRY|daten$STATUS),
barchart(table(daten$COUNTRY)|table(daten$STATUS)),