[R] Lattice: Understanding How Points Connected by Lines

[David Winsemius]

On Jan 1, 2012, at 5:50 PM, Rich Shepard wrote:

>  A data frame has one factor, one date, and one numeric column. When  
> I plot
> these using the default pch of the open circle (first attachment),
>
> xyplot(TDS ~ sampdate | she.s, data = sheep.cast, main = 'TDS in Sheep
> Creek', ylab = 'Concentration (mg/L)', xlab = 'Time')
>
> I see the higher concentration points toward the right on several  
> panels.
> But, when I change to using a line type (second attachment),
>
> xyplot(TDS ~ sampdate | she.s, data = sheep.cast, type = 'l', main =  
> 'TDS in
> Sheep Creek', ylab = 'Concentration (mg/L)', xlab = 'Time')
>
> the higher concentrations are connected to points far to the left  
> (earlier
> years). I expected to see a line with a few spikes toward the right  
> sides of
> the panels.
>
>  Where do I start to look for the reason for this behavior?

You should look at the structure of your dataset variables and the  
order in which they occur  ... none  of which is "visible" in what you  
have offered.

-- 

David Winsemius, MD
West Hartford, CT