COMPGROUPS.NET | Search | Post Question | Groups | Stream | About | Register

### test of spatial dependence??

• Follow

```hello SAS-friends,

I am a R beginner and try to ask a basic question:

How to test the spatial dependence of a column of data? for example, I have
25 agricultural fields, and I measure the average slope (%) or pH for each
field. All I have is 25 numbers.

PS, could someone confirm that "spatial dependence" is equivalent to
"spatial correlation" or "spatial autocorrelation" or not.

Thank you very much.
XY

--
say hello to friends of STAT and GIS
```
 0
Reply sasncsu (16) 12/5/2006 9:37:35 PM

```sasncsu@GMAIL.COM wrote:
>
>hello SAS-friends,
>
>I am a R beginner and try to ask a basic question:

If you are an R beginner, are you asking in the right place?
R has lots of spatial statistics capabilities, and a good mailing
list too.

That said, I'll stick to SAS here.

>How to test the spatial dependence of a column of data? for example, I have
>25 agricultural fields, and I measure the average slope (%) or pH for each
>field. All I have is 25 numbers.

Do you have distances between fields, or {x,y} coordinates?
If you don't have some capability of constructing the distances,
then there's no way to know which fields are 'close' to each
other and which are far apart.  So there would be no way to
do this determination.

>PS, could someone confirm that "spatial dependence" is equivalent to
>"spatial correlation" or "spatial autocorrelation" or not.

It depends on who asked the question.  Really.  'spatial
dependence' is a really vague term.  What are you planning
on using the information for?  Modeling spatial correlations
in some manner of linear model?

>Thank you very much.
>XY
>
>
>--
>say hello to friends of STAT and GIS

Take a look at PROC VARIOGRAM to do the modeling of the
spatial structure.

HTH,
David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330

_________________________________________________________________
Talk now to your Hotmail contacts with Windows Live Messenger.
http://clk.atdmt.com/MSN/go/msnnkwme0020000001msn/direct/01/?href=http://get.live.com/messenger/overview
```
 0
Reply davidlcassell (5578) 12/6/2006 5:46:01 AM

```Thanks for reply, David. No, I don't have the coordiates or the
distance. In other words, my data are not point data. But I think
there is a way to test of spatial dependence for areal data or lattice
data. In this case, the variable of interest is typically the average
value of an area instead of a point. Is a proximity matrix composed of
0 and 1 enough (1 means two units are adjacent and 0 not)? Do you how
to do this?

Thank you.
Xu

On 12/6/06, David L Cassell <davidlcassell@msn.com> wrote:
> sasncsu@GMAIL.COM wrote:
> >
> >hello SAS-friends,
> >
> >I am a R beginner and try to ask a basic question:
>
>
> If you are an R beginner, are you asking in the right place?
> R has lots of spatial statistics capabilities, and a good mailing
> list too.
>
> That said, I'll stick to SAS here.
>
> >How to test the spatial dependence of a column of data? for example, I have
> >25 agricultural fields, and I measure the average slope (%) or pH for each
> >field. All I have is 25 numbers.
>
> Do you have distances between fields, or {x,y} coordinates?
> If you don't have some capability of constructing the distances,
> then there's no way to know which fields are 'close' to each
> other and which are far apart.  So there would be no way to
> do this determination.
>
> >PS, could someone confirm that "spatial dependence" is equivalent to
> >"spatial correlation" or "spatial autocorrelation" or not.
>
> It depends on who asked the question.  Really.  'spatial
> dependence' is a really vague term.  What are you planning
> on using the information for?  Modeling spatial correlations
> in some manner of linear model?
>
> >Thank you very much.
> >XY
> >
> >
> >--
> >say hello to friends of STAT and GIS
>
> Take a look at PROC VARIOGRAM to do the modeling of the
> spatial structure.
>
> HTH,
> David
> --
> David L. Cassell
> mathematical statistician
> Design Pathways
> 3115 NW Norwood Pl.
> Corvallis OR 97330
>
> _________________________________________________________________
> Talk now to your Hotmail contacts with Windows Live Messenger.
> http://clk.atdmt.com/MSN/go/msnnkwme0020000001msn/direct/01/?href=http://get.live.com/messenger/overview
>

--
say hello to friends of STAT and GIS
```
 0
Reply sasncsu (16) 12/6/2006 6:31:58 PM

```I would recommend you read the monograph on spatial autocorrelation by
Daniel Griffith. It provides good explanations,examples and formulae so
you can code the appropriate solution for your problem in R or SAS or
whatever.

You don't necessarily need distance for spatial autocorrelation. What
you do need is data on the topology of your fields and then create some
type of adjacency matrix. The simplest is a binary matrix for adjacent
sides - if two fields share a common border, code 1, if they do not,
code 0.

For small problems, you can usually do this by hand, based on a map of
the area of interest. Larger problems usually require a suitable GIS
that can create such an adjacency matrix - ARCVIEW used to do this
specifically for calculating spatial statistics, but I am not sure it
still does. Once you have this matrix, the standard spatial correlations
are straightforward to code.

Hope this helps,

Frank

-----Original Message-----
From: David L Cassell [mailto:davidlcassell@MSN.COM]
Sent: Wednesday, December 06, 2006 12:46 AM
Subject: Re: test of spatial dependence??

sasncsu@GMAIL.COM wrote:
>
>hello SAS-friends,
>
>I am a R beginner and try to ask a basic question:

If you are an R beginner, are you asking in the right place?
R has lots of spatial statistics capabilities, and a good mailing list
too.

That said, I'll stick to SAS here.

>How to test the spatial dependence of a column of data? for example, I
>have
>25 agricultural fields, and I measure the average slope (%) or pH for
>each field. All I have is 25 numbers.

Do you have distances between fields, or {x,y} coordinates?
If you don't have some capability of constructing the distances, then
there's no way to know which fields are 'close' to each other and which
are far apart.  So there would be no way to do this determination.

>PS, could someone confirm that "spatial dependence" is equivalent to
>"spatial correlation" or "spatial autocorrelation" or not.

It depends on who asked the question.  Really.  'spatial dependence' is
a really vague term.  What are you planning on using the information
for?  Modeling spatial correlations in some manner of linear model?

>Thank you very much.
>XY
>
>
>--
>say hello to friends of STAT and GIS

Take a look at PROC VARIOGRAM to do the modeling of the spatial
structure.

HTH,
David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330

_________________________________________________________________
Talk now to your Hotmail contacts with Windows Live Messenger.
http://clk.atdmt.com/MSN/go/msnnkwme0020000001msn/direct/01/?href=http:/
/get.live.com/messenger/overview
```
 0
Reply FIvis (31) 12/6/2006 7:07:01 PM

3 Replies
41 Views