scatterplot matrix question

I would like a scatterplot matrix and a correlation matrix for the
following set-up.
The data (dataframe d) are like this:

angle  resp
-90      182
-60      137
-30      ...etc
0
30
60
90
...etc

I would like each cell in the matrix to be the scatterplot of the
responses for each pair of angles ( -90 vs -60, -90 vs -30, etc). Same
for the correlation matrix.

Please tell me what to do. Thanks very much!

Bill

On Sat, 2 Jan 2010, William Simpson wrote:

I would like a scatterplot matrix and a correlation matrix for the
following set-up.
The data (dataframe d) are like this:

angle  resp
-90      182
-60      137
-30      ...etc
0
30
60
90
...etc

I would like each cell in the matrix to be the scatterplot of the
responses for each pair of angles ( -90 vs -60, -90 vs -30, etc). Same
for the correlation matrix.

Please tell me what to do. Thanks very much!

1) You need to "provide commented, minimal, self-contained, reproducible
  code."

2) You need to fill in some missing info: Either you have only one
  response for each angle, or you need a third variable to pair up the
  corresponding responses for one angle with those of another.

On Sat, Jan 2, 2010 at 4:55 PM, Charles C. Berry <cberry at tajo.ucsd.edu> wrote:

On Sat, 2 Jan 2010, William Simpson wrote:

I would like a scatterplot matrix and a correlation matrix for the
following set-up.
The data (dataframe d) are like this:

angle  resp
-90      182
-60      137
-30      ...etc
0
30
60
90
...etc

I would like each cell in the matrix to be the scatterplot of the
responses for each pair of angles ( -90 vs -60, -90 vs -30, etc). Same
for the correlation matrix.

Please tell me what to do. Thanks very much!

1) You need to "provide commented, minimal, self-contained, reproducible
  code."

d<-read.table("rstuff/data.dat",header=TRUE)

Now what? :-)

certainly it's not
pairs(d)

2) You need to fill in some missing info: Either you have only one
  response for each angle, or you need a third variable to pair up the
  corresponding responses for one angle with those of another.

contents of rstuff/data.dat:

 angle  resp  ID
 -90      182  1
 -60      137  1
 -30      123  1
 0    67  1
 30   32  1
 60   12  1
 90    13  1
-90      178  2
 -60      111  2
 -30      137  2
 0    94  2
 30   59  2
 60   1  2
 90    19  2

I actually have a lot more than 2 experimental units (ID)...

Thanks for any help
Bill

Now that we are able to help with some more detailed view of your data
(although you could have helped helping by making it easier for us to import
your data into R), the answer is:

Almost, after reshaping:



dwide <- reshape(d, v.names="resp", idvar="ID",
                timevar="angle", direction="wide")
pairs(dwide[,-1])

Now that we are able to help with some more detailed view of your  
data
(although you could have helped helping by making it easier for us  
to import
your data into R), the answer is:

I am preparing for the data analysis, writing the code (knowing I may
have to modify it later) while the data are still being collected.
That's why I used the artificial data when asked for some. I will be
pressed for time when the data arrive.

Almost, after reshaping:



dwide <- reshape(d, v.names="resp", idvar="ID",
               timevar="angle", direction="wide")
pairs(dwide[,-1])

Thanks very much, Uwe. I will try this (on artificial data). I think
reshape() requires a library [reshape?].

Thanks very much, Uwe. I will try this (on artificial data). I think
reshape() requires a library [reshape?].

No. In fact, the reshape package does not have a reshape function.

--

David Winsemius, MD
Heritage Laboratories
West Hartford, CT