ggplot2 syntax changes

[ropensci]

ggplot2 syntax changes

[karthik]

Also, it would be great to allow people to manipulate the plot the way they choose. So add a ... argument to the outer function and pass it on to ggplot.

[sckott]

Pointtype and jitter fixed.

The problem I'm having with passing options into the function is that I don't want the options passed in to override all of the options we specify within the function. For example, in our block of code for plotting:

ggplot(world, aes(long, lat)) + # make the plot
    geom_polygon(aes(group=group), fill="white", color="gray40", size=0.2) +
    geom(data=tomap, aes(decimalLongitude, decimalLatitude, colour=taxonName), 
         alpha=0.4, size=3, position=jitter) +
    labs(x="", y="") +
    theme_bw(base_size=14)

where would we allow users to pass in different options? We could let people do something like specify a set of ggplot options in a function, then pass it in to gbifmap

mycustom <- function(){
    list(geom_point(size=9), 
              etc....)
}
out <- occurrencelist(scientificname = 'Accipiter erythronemius', coordinatestatus = TRUE, maxresults = 100)
gbifmap( out,  customize = mycustom() )

And then modify our above plotting block to add whatever options the user wants

ggplot(world, aes(long, lat)) + # make the plot
    geom_polygon(aes(group=group), fill="white", color="gray40", size=0.2) +
    geom(data=tomap, aes(decimalLongitude, decimalLatitude, colour=taxonName), 
         alpha=0.4, size=3, position=jitter) +
    labs(x="", y="") +
    theme_bw(base_size=14) +
    customize

But then the user options override the previous options in the ggplot code block

[karthik]

I just realized that this is a non-issue. Once a ggplot2 object is returned, anyone can override your settings or add more. For example, I could do a + geom_point(size = 14) and it will override the one inside. So we can just tell users that they can customize however they like after the fact (no need to pass arguments via the function call).

[sckott]

not sure that's true. If you do geom_point(size = 14) then the geom(data=tomap, aes(decimalLongitude, decimalLatitude, colour=taxonName), alpha=0.4, size=3, position=jitter) code is ignored, and screws up the plot.

[karthik]

Hmm. See this question for reference: http://stackoverflow.com/questions/13831717/how-does-one-pass-additional-parameters-to-a-ggplot-generated-inside-a-function/13832162#13832162

Perhaps we should work on this some more to make the theme part general but not break core functionality.

On Wednesday, December 12, 2012 at 11:47 AM, Scott Chamberlain wrote:

not sure that's true. If you do geom_point(size = 14) then the geom(data=tomap, aes(decimalLongitude, decimalLatitude, colour=taxonName), alpha=0.4, size=3, position=jitter) code is ignored, and screws up the plot.

—
Reply to this email directly or view it on GitHub (#11 (comment)).

[karthik]

Another bug, the legend should show up even if there is only one species. Don't you agree? If so I can code that fix.

[sckott]

I tend to not think so, but perhaps it makes sense for greater consistency in map output

[karthik]

scratch that. You are right. plots with only one type don't need legends.

On Wednesday, December 12, 2012 at 2:18 PM, Scott Chamberlain wrote:

I tend to not think so, but perhaps it makes sense for greater consistency in map output

—
Reply to this email directly or view it on GitHub (#11 (comment)).

[sckott]

okay, cool

[sckott]

think this is done