R help - Jul 2009 - OK - I got the data - now what? :-)

OK, I guess I'm getting better at the data part of R. I wrote a
program outside of R this morning to dump a bunch of experimental
data. It's a sort of ragged array - about 700 rows and 400 columns,
but the amount of data in each column varies based on the length of
the experiment. The real data ends with a 0 following some non-zero
value. It might be as short as 5 to 10 columns or as many as 390. The
first 9 columns contain some data about when the experiment was run
and a few other things I thought I might be interested in later. All
the data starts in column 10 and has headers saying C1, C2, C3, C4,
etc., up to C390 The first value for every experiment is some value I
will normalize and then the values following are above and below the
original tracing out the path that the experiment took, ending
somewhere to the right but not a fixed number of readings.

R reads it in fine and it looks good so far.

Now, what I thought I might do with R is plot all 700 rows as
individual lines, giving them some color based on info in columns 1-9,
but suddenly I'm lost again in plots which I think should be fairly
easy. How would I go about creating a plot for even one line, much
less all of them? I don't have a row with 1,2,3,4 to us as the X axis
values. I could go back and put one in the data but then I don't think
that should really be required, or I could go back and make the
headers for the whole array 1:400 and then plot from 10:400 but I
thought I read that headers cannot start with numbers.

Maybe the X axis values for a plot can actually be non-numeric C1, C2,
C3, C4, etc and I could use line (C1,0) to (C2,5) and so on? Or maybe
I should strip the C from C1 and be left with 1? Maybe the best thing
is to copy the data for one line to another data.frame or array and
then plot that?

Just sort of lost looking at help files. Thanks for any ideas you can
send along. Ask questions if I didn't explain my problem well enough.
Not looking for anyone to do my work, just trying to get the concepts
right

Cheers,
Mark

See if this example helps; show how to either plot the row or columns
of a data frame:
> test <- data.frame(C1=runif(10), C2=runif(10), C3=runif(10))
> test
           C1        C2        C3
1  0.91287592 0.3390729 0.4346595
2  0.29360337 0.8394404 0.7125147
3  0.45906573 0.3466835 0.3999944
4  0.33239467 0.3337749 0.3253522
5  0.65087047 0.4763512 0.7570871
6  0.25801678 0.8921983 0.2026923
7  0.47854525 0.8643395 0.7111212
8  0.76631067 0.3899895 0.1216919
9  0.08424691 0.7773207 0.2454885
10 0.87532133 0.9606180 0.1433044
> # this will plot each column (C1, C2, C3)
> matplot(test, type='o')
> # plot each row
> matplot(t(test), type='o')

On Sat, Jul 4, 2009 at 8:02 PM, Mark Knecht<markknecht at gmail.com>
wrote:
> OK, I guess I'm getting better at the data part of R. I wrote a
> program outside of R this morning to dump a bunch of experimental
> data. It's a sort of ragged array - about 700 rows and 400 columns,
> but the amount of data in each column varies based on the length of
> the experiment. The real data ends with a 0 following some non-zero
> value. It might be as short as 5 to 10 columns or as many as 390. The
> first 9 columns contain some data about when the experiment was run
> and a few other things I thought I might be interested in later. All
> the data starts in column 10 and has headers saying C1, C2, C3, C4,
> etc., up to C390 The first value for every experiment is some value I
> will normalize and then the values following are above and below the
> original tracing out the path that the experiment took, ending
> somewhere to the right but not a fixed number of readings.
>
> R reads it in fine and it looks good so far.
>
> Now, what I thought I might do with R is plot all 700 rows as
> individual lines, giving them some color based on info in columns 1-9,
> but suddenly I'm lost again in plots which I think should be fairly
> easy. How would I go about creating a plot for even one line, much
> less all of them? I don't have a row with 1,2,3,4 to us as the X axis
> values. I could go back and put one in the data but then I don't think
> that should really be required, or I could go back and make the
> headers for the whole array 1:400 and then plot from 10:400 but I
> thought I read that headers cannot start with numbers.
>
> Maybe the X axis values for a plot can actually be non-numeric C1, C2,
> C3, C4, etc and I could use line (C1,0) to (C2,5) and so on? Or maybe
> I should strip the C from C1 and be left with 1? Maybe the best thing
> is to copy the data for one line to another data.frame or array and
> then plot that?
>
> Just sort of lost looking at help files. Thanks for any ideas you can
> send along. Ask questions if I didn't explain my problem well enough.
> Not looking for anyone to do my work, just trying to get the concepts
> right
>
> Cheers,
> Mark
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
Jim Holtman
Cincinnati, OH
+1 513 646 9390

What is the problem that you are trying to solve?

Hi. Essentially your data is currently in "wide" format, with repeated
measures in different columns. For most analysis and in particular for
graphing, it is frequently helpful to reshape your data into a "long"
format, with one row per data value and additional variables to list
experiment or subject identifier, experimental conditions etc.

see ?reshape and Dr. Wickham's reshape package (http://had.co.nz/reshape/)

Good luck,

Mark


2009/7/5 Mark Knecht <markknecht at gmail.com>:
> OK, I guess I'm getting better at the data part of R. I wrote a
> program outside of R this morning to dump a bunch of experimental
> data. It's a sort of ragged array - about 700 rows and 400 columns,
> but the amount of data in each column varies based on the length of
> the experiment. The real data ends with a 0 following some non-zero
> value. It might be as short as 5 to 10 columns or as many as 390. The
> first 9 columns contain some data about when the experiment was run
> and a few other things I thought I might be interested in later. All
> the data starts in column 10 and has headers saying C1, C2, C3, C4,
> etc., up to C390 The first value for every experiment is some value I
> will normalize and then the values following are above and below the
> original tracing out the path that the experiment took, ending
> somewhere to the right but not a fixed number of readings.
>
> R reads it in fine and it looks good so far.
>
> Now, what I thought I might do with R is plot all 700 rows as
> individual lines, giving them some color based on info in columns 1-9,
> but suddenly I'm lost again in plots which I think should be fairly
> easy. How would I go about creating a plot for even one line, much
> less all of them? I don't have a row with 1,2,3,4 to us as the X axis
> values. I could go back and put one in the data but then I don't think
> that should really be required, or I could go back and make the
> headers for the whole array 1:400 and then plot from 10:400 but I
> thought I read that headers cannot start with numbers.
>
> Maybe the X axis values for a plot can actually be non-numeric C1, C2,
> C3, C4, etc and I could use line (C1,0) to (C2,5) and so on? Or maybe
> I should strip the C from C1 and be left with 1? Maybe the best thing
> is to copy the data for one line to another data.frame or array and
> then plot that?
>
> Just sort of lost looking at help files. Thanks for any ideas you can
> send along. Ask questions if I didn't explain my problem well enough.
> Not looking for anyone to do my work, just trying to get the concepts
> right
>
> Cheers,
> Mark
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
>


-- 
Dr. Mark Wardle
Specialist registrar, Neurology
Cardiff, UK