[R] Bivariate Scatter Plots with Lattice

[Rich Shepard]

On Thu, 22 Sep 2011, Dennis Murphy wrote:

> Question: Do you want 37 different panels with plots of quant vs. date
> by param, or two panels (one per chemical) with all 37 streams? If you
> only want two of the eight chemicals,

Dennis,

   I want only 2 chemicals per plot.

> I'd suggest using subset() to select out the pair you want and then
> redefine the param factor so that the subset data frame has two factor
> levels instead of eight.

   Sure. I keep working with the original data set and haven't yet fully
learned to create subsets as needed.

> Since there are several ways to interpret 'in which there are data', you
> might want to expound a little further about your intentions in that
> regard.

   What I mean that over all locations and chemicals the earliest date for
which there is data is 1980-03-01. However, some chemicals might not have
been required to be monitored before, say, 1990. If the pair in a given plot
have no data prior to 1990 (for example), that's what I want as the xlim.

Thanks,

Rich