Skip to content

SaTScan in R

6 messages · Kitty Lee, Barry Rowlingson, Renaud Lancelot +2 more

#
I'm wondering is there a package in R that has similar functions as SaTscan? In particular, I'm looking for space-time cluster detection. The D-cluster package doesn't appear to have a function for space-time detection (please correct me if I'm wrong).

Any ideas would be highly appreciated!


K.
#
On Fri, Dec 3, 2010 at 8:25 AM, Kitty Lee <lee.kitty at yahoo.com> wrote:
SaTScan (TM) seems to be a Java GUI on a whole bunch of scan
statistics devised by Kulldorff. Running RSiteSearch("kulldorff")
doesn't find any implementations of space-time scan statistics in R.
So I'm thinking not.

 But it would be a heck of a great project to start. SaTScan (TM) is
closed source, but has a fairly permissive license (it allows
redistribution). I thought it would be able to produce pretty maps
like on its web page, but that appears to be misleading - one would
have to load the output data into a GIS (or R) for mapping.

 Also, don't forget that SaTScan(TM) is trademarked to Kulldorff so
you have to say (TM) every time you mention it. All the methods are
published, so reimplementations of them should not be covered by
trademark, patent, copyright problems (oh, I'm not a lawyer...).

 So basically, the answer is no, but if Kulldorff reads this, how
about opening the source?

Barry
#
Le 03/12/2010 18:57, Barry Rowlingson a ?crit :
That would be great! I feel we miss these methods in the current 
implementation of R packages for spatial epidemiology.

All the best,

Renaud

  
    
1 day later
9 days later
#
You may find R codes for space-time cluster detection in Tango's
"Statistical Methods for Disease Clustering"
(http://www.amazon.com/dp/1441915710/). I do not write here the
downloadable url of these R functions written in the book because I am
not sure it can be disclosed to public. So, what I can advise to you
is that you may read the book.

I hope it can help you.

--
Susumu

On Fri, 3 Dec 2010 00:25:40 -0800 (PST)
Kitty Lee <lee.kitty at yahoo.com> wrote:

            
#
Kitty,

You might want to try SpatialEpi (although I think it only has Spatial, not Spatio-Temporal detection)

cheers

Ben
On 03/12/2010, at 7:25 PM, Kitty Lee wrote: