[R] Fwd: [R-sig-hpc] Holtman's levelplot ?

[David Winsemius]

On Aug 24, 2010, at 2:56 PM, Michel Lutz wrote:

> Hello,
>
> As recommended by N. Gunther, I'm writing you because I have some
> difficulties to create a script to track computer performance, as  
> done in
> Holtman, 2005 (Visualization techniques for analysing patterns in  
> system
> performance).
>
> I would like to do a leveplot, with the kind of file here enclosed.

Nope. You probably did not read all the way to the end of the Posting  
Guide to see what kinds of attachments are allowed on this list.

>
> I assume I should create a matrix in R, then use the function 'image'.

Maybe. There is an image function for base graphics:

?image

First example is very similar to the first example in levelplot. They  
have very similar data layout requirements.

>
> As I am not familiar with the lattice package, I have some  
> difficulties to
> move ahead : I think my pb is in the creation of the matrix. I am  
> not able
> to create sthing consistent with the image function. I am even  
> thinking I
> should build my input csv file differently, but I am not sure how to  
> do...
>
> Is someone able to help me ?

Given Jim's propensity to finish his emails with "What problem are you  
trying to solve", let me beat him to the punch. What's you data layout  
look like and what do you want the plot to look like? If you can  
convert that file to plain text and make sure its file extension  
is .txt, it will probably make it through if you resend it.

You can get some other ideas for specifics by running the first   
example on the levelplot help page:

?levelplot

Lattice graphic do not mix well with base graphics so you need to keep  
them segregated in your plotting kitchen if you want to remain r-kosher.

-- 
David.
>
> Thank you very much !
>
> REgards,
>
> Michel
>
>
>
> ---------- Forwarded message ----------
> From: Red Roo <redrooz at yahoo.com>
> Date: Tue, Aug 24, 2010 at 3:46 PM
> Subject: Re: [R-sig-hpc] Holtman's levelplot ?
> To: Michel Lutz <miclutz at gmail.com>
> Cc: "r-sig-hpc at r-project.org" <r-sig-hpc at r-project.org>, redrooz at yahoo.com
>
>
> Michel,
>
> If you ask your plotting questions over at r-help at r-project.org, Jim  
> Holtman
> will probably address them directly. He is also an instructor for my
> "Guerrilla
> Data Analysis Techniques" class
> http://www.perfdynamics.com/Classes/Outlines/gdata.html and we just
> discussed
> this plot 2 weeks ago.
>
>
> --Neil Gunther
>
> Taking The Pith Out Of Performance http://perfdynamics.blogspot.com/
> Follow me on Twitter http://twitter.com/DrQz
> Performance Dynamics Company http://www.perfdynamics.com/
>
>
>>>>
>>>
>>> On 08/24/2010 03:47 AM, Michel Lutz wrote:
>>>
>>> Hello to all,
>>>>
>>>> I am sorry for disturbing you, but I really don't know where/who to
> ask...
>>>> so if this distribution list is not dealing with my concern, please
> excuse
>>>> me...
>>>>
>>>> Here is my concern : I am trying to do a levelplot as in Holtman,  
>>>> 2005
>>>> (Visualization techniques for analysing patterns in system  
>>>> performance
>>>> data), to display CPU usage according to the day of the month (y)  
>>>> and
>>>> minutes (x).
>>>>
>>>> I am familiar with R basic features, but not at all with lattice  
>>>> package
> and
>>>> I don't understand the logic of this leveplot function,  
>>>> especially in the
>>>> particular case of a time series.
>>>>
>>>> I've read all the material found in the web, but I am completely  
>>>> stuck...
>>>>
>>>> Does someone have some clues or documentation that could help me  
>>>> to draw
>>>> such a plot ? (which is nothing new : the plot I'm dreaming of is  
>>>> the
>>>> Holtman's levelplot !).
>>>>
>>>> Thank you very much, and excuse-me if I'm not addressing the right
>>>> distribution list.
>>>>
>>>> Best regards,
>>>>
>>>> Michel Lutz
>>>> France - PhD Student

David Winsemius, MD
West Hartford, CT