14 Nov
2010
14 Nov
'10
8:49 a.m.
Hi all,
Basically I'm getting an error message that I don't really understand. It
says that the subscript out of bounds but I can't see how.
Thank you for your help,
Best,
Antonio.
# missing date matrix
k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
k1 <-as.matrix(k1)
k1 <-as.matrix(k1)
str(k1)
num [1:4126, 1:3] 160 NA NA NA NA ...
dim(k1)
[1] 4126 3
head(k1)
[,1] [,2] [,3]
[1,] 160 1962 1
[2,] NA 1963 1
[3,] NA 1964 1
[4,] NA 1965 1
[5,] NA 1966 1
[6,] 150 1967 1
quartz()
missmap(k1,csvar=k1[,3],tsvar=k1[,2])
Error in missmap(k1, csvar = k1[, 3], tsvar =
k1[, 2]) :
subscript out of bounds
15 Nov
15 Nov
4:46 a.m.
Antonio,
Instead of:
missmap(k1,csvar=k1[,3],tsvar=k1[,2])
See if:
missmap(k1,csvar=3,tsvar=2)
works in your
dataset. That is, the arguments are the column positions, not
the actual extracted columns from the matrix. Or if this were a data frame you
could use variable names, such as:
missmap(k1,csvar="countryname",tsvar="year")
if those names were
appropriate (that is, contained in "names(k1)")
Hope all is well with you,
James
On Sun, Nov 14, 2010 08:49 AM "Antonio P. Ramos"
<ramos.grad.student(a)gmail.com>
wrote:
>
>Hi all,
>
>Basically I'm getting an error message that I don't really understand. It
>says that the subscript out of bounds but I can't see how.
>
>Thank you for your help,
>
>Best,
>
>Antonio.
>
>
>> # missing date matrix
>>
>> k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
>> k1 <-as.matrix(k1)
>> k1 <-as.matrix(k1)
>> str(k1)
> num [1:4126, 1:3] 160 NA NA NA NA ...
>> dim(k1)
>[1] 4126 3
>> head(k1)
> [,1] [,2] [,3]
>[1,] 160 1962 1
>[2,] NA 1963 1
>[3,] NA 1964 1
>[4,] NA 1965 1
>[5,] NA 1966 1
>[6,] 150 1967 1
>> quartz()
> missmap(k1,csvar=k1[,3],tsvar=k1[,2])
>Error
in missmap(k1, csvar = k1[, 3], tsvar = k1[, 2]) :
> subscript out of bounds
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
5:39 a.m.
Hey James! How are you? Such a long time. We miss you!
Thank for your reply: basically using your tip everything run but the map is
kind of weird: I think it is plotting the 3 variables of the my date set,
when I actually care just about the first one. Thus I dunno how to select
the variable I really one see. If try this I get an error message:
quartz()
missmap(k1[,1],csvar=3,tsvar=2)
Error in colMeans(is.na(obj)) :
'x' must be an array of at least two dimensions
If I do the following code I get the attached graph, which is really the
thing I wanna plot:
k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
k1 <-as.data.frame(k1)
k1 <- as.matrix(k1)
str(k1)
num [1:4126, 1:3] 160 NA NA NA NA ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:3] "V1" "V2" "V3"
dim(k1)
[1] 4126 3
head(k1)
V1 V2 V3
[1,] 160 1962 1
[2,] NA 1963 1
[3,] NA 1964 1
[4,] NA 1965 1
[5,] NA 1966 1
[6,] 150 1967 1
quartz()
missmap(k1,csvar=3,tsvar=2)
Thank your for your time!
9:29 a.m.
Hi Antonio,
Unfortunately, you need to include the time and cross-sectional
variables in the data passed to missmap and they will be mapped along
with any other variables. The missmap function relies on the image()
function at its core, so you could try to look at the source and hack
together a map that suits your needs.
If you want to draw the map without any time/cross-section
information, you can force R to treat the column as a one-column
matrix:
missmap(k1[ ,1, drop = FALSE])
Hope that helps,
matt.
On Mon, Nov 15, 2010 at 5:39 AM, Antonio P. Ramos
<ramos.grad.student(a)gmail.com> wrote:
Hey James! How are you? Such a long time. We miss
you!
Thank for your reply: basically using your tip everything run but the map is
kind of weird: I think it is plotting the 3 variables of the my date set,
when I actually care just about the first one. Thus I dunno how to select
the variable I really one see. If try this I get an error message:
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
quartz()
missmap(k1[,1],csvar=3,tsvar=2)
Error in colMeans(is.na(obj)) :
'x' must be an array of at least two dimensions
If I do the following code I get the attached graph, which is really the
thing I wanna plot:
k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
k1 <-as.data.frame(k1)
k1 <- as.matrix(k1)
str(k1)
num [1:4126, 1:3] 160 NA NA NA NA ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:3] "V1" "V2" "V3"
dim(k1)
[1] 4126 3
head(k1)
V1 V2 V3
[1,] 160 1962 1
[2,] NA 1963 1
[3,] NA 1964 1
[4,] NA 1965 1
[5,] NA 1966 1
[6,] 150 1967 1
quartz()
missmap(k1,csvar=3,tsvar=2)
Thank your for your time!
6 p.m.
Thanks for your reply. I got it now.
I was just wondering about some of the common "problems" with these maps:
often in time-serious cross-sectional data we have so many counties that we
can't really plot all their names; and if we wanna arrange country in terms
of, say, income per capita income, to see the correlation of missing data
and this indicator, than is even more difficult. I saw that you did some
tricks in this post,like plotting over country names over both sides of the
y-axis but I can't see the code. Perhaps you and others might wanna share
some of this tricks.
http://www.iq.harvard.edu/blog/sss/archives/2009/02/missingness_map.shtml
Congratulation for your software!
Best,
Antonio.
On Mon, Nov 15, 2010 at 6:29 AM, Matt Blackwell <blackwel(a)fas.harvard.edu>wrote;wrote:
Hi Antonio,
Unfortunately, you need to include the time and cross-sectional
variables in the data passed to missmap and they will be mapped along
with any other variables. The missmap function relies on the image()
function at its core, so you could try to look at the source and hack
together a map that suits your needs.
If you want to draw the map without any time/cross-section
information, you can force R to treat the column as a one-column
matrix:
missmap(k1[ ,1, drop = FALSE])
Hope that helps,
matt.
On Mon, Nov 15, 2010 at 5:39 AM, Antonio P. Ramos
<ramos.grad.student(a)gmail.com> wrote:
Hey James! How are you? Such a long time. We miss
you!
Thank for your reply: basically using your tip everything run but the map
is
kind of weird: I think it is plotting the 3
variables of the my date set,
when I actually care just about the first one. Thus I dunno how to select
the variable I really one see. If try this I get an error message:
quartz()
missmap(k1[,1],csvar=3,tsvar=2)
Error in colMeans(is.na(obj)) :
'x' must be an array of at least two dimensions
If I do the following code I get the attached graph, which is really the
thing I wanna plot:
k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
k1 <-as.data.frame(k1)
k1 <- as.matrix(k1)
str(k1)
num [1:4126, 1:3] 160 NA NA NA NA ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:3] "V1" "V2" "V3"
dim(k1)
[1] 4126 3
head(k1)
V1 V2 V3
[1,] 160 1962 1
[2,] NA 1963 1
[3,] NA 1964 1
[4,] NA 1965 1
[5,] NA 1966 1
[6,] 150 1967 1
quartz()
missmap(k1,csvar=3,tsvar=2)
Thank your for your time!
16 Nov
16 Nov
9:56 a.m.
Hi Antonio,
Thanks for these suggestions. The next time we update the missmap
function, we will definitely keep these in mind. You can reorder the
map currently by reordering the matrix or data frame you pass to it.
So, for income you could do:
missmap(data[order(data$income),])
Cheers,
matt.
On Mon, Nov 15, 2010 at 6:00 PM, Antonio P. Ramos
<ramos.grad.student(a)gmail.com> wrote:
Thanks for your reply. I got it now.
I was just wondering about some of the common "problems" with these maps:
often in time-serious cross-sectional data we have so many counties that we
can't really plot all their names; and if we wanna arrange country in terms
of, say, income per capita income, to see the correlation of missing data
and this indicator, than is even more difficult. I saw that you did some
tricks in this post,like plotting over country names over both sides of the
y-axis but I can't see the code. Perhaps you and others might wanna share
some of this tricks.
http://www.iq.harvard.edu/blog/sss/archives/2009/02/missingness_map.shtml
Congratulation for your software!
Best,
Antonio.
On Mon, Nov 15, 2010 at 6:29 AM, Matt Blackwell <blackwel(a)fas.harvard.edu>
wrote:
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
Hi Antonio,
Unfortunately, you need to include the time and cross-sectional
variables in the data passed to missmap and they will be mapped along
with any other variables. The missmap function relies on the image()
function at its core, so you could try to look at the source and hack
together a map that suits your needs.
If you want to draw the map without any time/cross-section
information, you can force R to treat the column as a one-column
matrix:
missmap(k1[ ,1, drop = FALSE])
Hope that helps,
matt.
On Mon, Nov 15, 2010 at 5:39 AM, Antonio P. Ramos
<ramos.grad.student(a)gmail.com> wrote:
> Hey James! How are you? Such a long time. We miss you!
>
> Thank for your reply: basically using your tip everything run but the
> map is
> kind of weird: I think it is plotting the 3 variables of the my date
> set,
> when I actually care just about the first one. Thus I dunno how to
> select
> the variable I really one see. If try this I get an error message:
>
>> quartz()
>> missmap(k1[,1],csvar=3,tsvar=2)
> Error in colMeans(is.na(obj)) :
> 'x' must be an array of at least two dimensions
>
> If I do the following code I get the attached graph, which is really the
> thing I wanna plot:
>
> k1 <-cbind(data1$INFMORT,data1$YEAR,data1$COUNTRY)
>> k1 <-as.data.frame(k1)
>> k1 <- as.matrix(k1)
>> str(k1)
> num [1:4126, 1:3] 160 NA NA NA NA ...
> - attr(*, "dimnames")=List of 2
> ..$ : NULL
> ..$ : chr [1:3] "V1" "V2" "V3"
>> dim(k1)
> [1] 4126 3
>> head(k1)
> V1 V2 V3
> [1,] 160 1962 1
> [2,] NA 1963 1
> [3,] NA 1964 1
> [4,] NA 1965 1
> [5,] NA 1966 1
> [6,] 150 1967 1
>> quartz()
>> missmap(k1,csvar=3,tsvar=2)
>
> Thank your for your time!
>
4915
days inactive
4917
days old
amelia@lists.gking.harvard.edu
5 comments
3 participants
participants (3)
-
Antonio P. Ramos -
JAMES HONAKER -
Matt Blackwell