[R-sig-Geo] Creating density heatmaps for geographical data

[Roger Bivand]

On Wed, 20 Oct 2010, Rolf Turner wrote:

>
> The spatstat package is designed for the analysis of spatial point
> patterns.  In this context the existence of a window --- the
> *observation* window --- is absolutely crucial.  You have to know where
> points have been *looked for*, because there is information in where
> the points aren't, as well as in where they are.  And you can't say
> anything about where they aren't unless you know where they have been
> looked for.
>
> If you are putting spatstat to other uses where the nature of the process
> generating the points is not of interest, and you want a default window
> which has the nature of a bounding box, then this is easy to arrange.
> E.g. you could write your own simple little utility, say
>
> ppp_koh <- function(x,y,wrecked=TRUE,f,marks=NULL) {
> require(spatstat)
> ppp(x,y,window=ripras(x,y,shape=if(wrecked) "rectangle" else "convex",f=f),marks=marks)
> }
>
> Then you can do things like:
> set.seed(42)
> x <- runif(100,3,7)
> y <- runif(100,-2,9)
> X <- ppp_koh(x,y)

Or using the suite of coercion methods in in the maptools package:

library(maptools)
set.seed(42)
x <- runif(100,3,7)
y <- runif(100,-2,9)
xy <- SpatialPoints(cbind(x, y))
X <- as(xy, "ppp")

and so on. There are horses for courses, and the ppp, psp, im, and other 
representations, like polygons windows, are well tailored to their 
applications. So using coercion methods is a less invasive way of 
providing interoperability than trying to make a single set of 
representations fit all uses.

Roger

>
> Note that by default the ripras() function called by ppp_koh() expands the
> bounding box of the points slightly from extent of the data (to give a ``maximum
> likelihood estimate'' of the bounding box).
>
> If you don't want this expansion then specify f=1:
>
> Y <- ppp_koh(x,y,f=1)
>
> Note that this will cause at least two points of the pattern (and probably
> four or more) to lie on the boundary of the window.
>
> In genuine spatial point process applications one is advised to eschew
> the use of the ripras() function.
>
> 	cheers,
>
> 		Rolf
>
> On 19/10/2010, at 11:58 PM, Karl Ove Hufthammer wrote:
>
>> Rolf Turner wrote:
>>
>>>> 'ppp' objects used by 'spatstat' have an annoying structure
>>>
>>> <SNIP>
>>>
>>> Annoyance is in the mind of the beholder I think.  Personally I
>>> find the structure of ppp objects absolutely *delightful*! :-)
>>>
>>> Also, they are simple, intuitive, easy to work with, and easy to
>>> understand, take apart, dissect, and manipulate.  Furthermore the
>>> tools in spatstat make it easy to do unto ppp objects pretty well
>>> anything one might want to do to a point pattern.  (I'm not being
>>> immodest here; the tools are mainly down to Adrian.)
>>>
>>> Just what do you find ``annoying'' about the structure of ppp objects?
>>
>> Well, I guess only the fact that they’re not ‘sp’ objects, really. And
>> that’s very annoying … :)
>>
>> And though the ‘window’ element of ‘ppp’ objects may be of use to some
>> people, I haven’t had any use for it. The annoying thing here is that the
>> constructor doesn’t generate the window automatically, based on the extent /
>> bounding box of the data, and don’t have an *option* for doing this, either.
>> Whenever I have used ‘spatstat’ (not too often), I have had to spend too
>> much time looking up how the window should be specified. Having [0,1] ×
>> [0,1] as the *default* window, and excluding any points outside this does
>> seems like a strange design decision.
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no