[R] Reshape or Plyr?

[arun]

Hi Bruce,
May be this helps.

library(plyr)

 res<-mutate(ddply(dat1,.(SPEC_CODE,Location_name),function(x) colSums(x[,-c(1:3,6)])),RA=10*(AI/Survey_Time))
 res
#   SPEC_CODE    Location_name Survey_Time AI         RA
#1       Buzz 079-f2p1-Acetuna        72.8  8  1.0989011
#2     Eumspp 079-f2p1-Acetuna        24.3  5  2.0576132
#3       Frag 079-f2p1-Acetuna        12.1 18 14.8760331
#4     Molmol 079-f2p1-Acetuna        12.1  1  0.8264463
#5     Molspp 079-f2p1-Acetuna        72.8 28  3.8461538
#6     Myokea 079-f2p1-Acetuna        12.2  1  0.8196721
#7     Nocalb 079-f2p1-Acetuna        24.3 10  4.1152263
#8     Phyllo 079-f2p1-Acetuna        36.4  4  1.0989011
#9     Ptedav 079-f2p1-Acetuna        36.4  3  0.8241758
#10    Ptegym 079-f2p1-Acetuna        36.4  6  1.6483516
#11    Ptepar 079-f2p1-Acetuna        36.4  9  2.4725275
#12    Pteper 079-f2p1-Acetuna        24.3  4  1.6460905
#13    Rhotum 079-f2p1-Acetuna        36.4 30  8.2417582
#14    Sacbil 079-f2p1-Acetuna        36.4 11  3.0219780
#15    Saclep 079-f2p1-Acetuna        36.4 32  8.7912088


write.csv(res,file="BruceFile.csv",row.names=FALSE)

A.K.


----- Original Message -----
From: Bruce Miller <batsncats at gmail.com>
To: arun <smartpink111 at yahoo.com>
Cc: 
Sent: Sunday, April 21, 2013 6:18 PM
Subject: Re: [R] Reshape or Plyr?

Hi Arun,

Your first example is what looks like the correct one but needs the 
Location_name as in your 2nd example.
I do want to group by "Location_name". The full data set will include up 
to 100 Location_names so having this along with SPEC_CODE, Survey_Time, 
AI and RA is great.
Then I need to add the line to write the results to a CSV or tab 
delimited file.

No need for the Start_Day as in your 2nd example.

On 4/21/2013 5:38 PM, arun wrote:
> Hi Bruce,
>
>  From your second email, I think you solved the problem.  But, still there is confusion due to the wordings.
> res1<-mutate(ddply(dat1,.(SPEC_CODE),function(x) colSums(x[,-c(1:3,6)])),RA=10*(AI/Survey_Time))
>   res1
>     SPEC_CODE Survey_Time AI         RA
> 1       Buzz        72.8  8  1.0989011
> 2     Eumspp        24.3  5  2.0576132
> 3       Frag        12.1 18 14.8760331
> 4     Molmol        12.1  1  0.8264463
> 5     Molspp        72.8 28  3.8461538
> 6     Myokea        12.2  1  0.8196721
> 7     Nocalb        24.3 10  4.1152263
> 8     Phyllo        36.4  4  1.0989011
> 9     Ptedav        36.4  3  0.8241758
> 10    Ptegym        36.4  6  1.6483516
> 11    Ptepar        36.4  9  2.4725275
> 12    Pteper        24.3  4  1.6460905
> 13    Rhotum        36.4 30  8.2417582
> 14    Sacbil        36.4 11  3.0219780
> 15    Saclep        36.4 32  8.7912088
>
>
> #or
> dat1$RA<-10*with(dat1,ave(AI,SPEC_CODE,FUN=sum))/with(dat1,ave(Survey_Time,SPEC_CODE,FUN=sum))
> head(dat1)
> #    Location_name SPEC_CODE Start_Day Survey_Time AI     Std_AI        RA
> #1 079-f2p1-Acetuna      Buzz 2/14/2012        12.1  1  0.8264463  1.098901
> #2 079-f2p1-Acetuna      Buzz 2/14/2012        12.1  1  0.8264463  1.098901
> #3 079-f2p1-Acetuna    Eumspp 2/14/2012        12.1  1  0.8264463  2.057613
> #4 079-f2p1-Acetuna      Frag 2/14/2012        12.1 18 14.8760300 14.876033
> #5 079-f2p1-Acetuna    Molspp 2/14/2012        12.1  5  4.1322310  3.846154
> #6 079-f2p1-Acetuna    Molspp 2/14/2012        12.1  5  4.1322310  3.846154
>
>
>
> ________________________________
>   From: Bruce Miller <batsncats at gmail.com>
> To: arun <smartpink111 at yahoo.com>
> Sent: Sunday, April 21, 2013 4:19 PM
> Subject: Re: [R] Reshape or Plyr?
>  
>
>
> Hi Arun,
>
> This looks good; however the not quite correct yet.
>
> Looks like the mutate option may be a good place to start.
> Sorry I did not clearly explain this in my original query. For
>        adjustment by unit effort I need the "Base AI per hour" summed for
>        the location.  Once I have that it can be standardized by 10
>        hours. An example for a single from your output
>
> #9     Ptedav 079-f2p1-Acetuna        36.4  3   12.1333333       121.333333
> Is not the correct way to determine AI per hour. So total survey hours for a 3 nights is 36.4 then the AI value is 3 so the AI per hour for Ptedav would be3/36.4 = .08241 then standardized for 10 hours of survey time would be 0.824175824.
> Now I need to determine how to re-arrange your code to reflect the
>        Summed AI/Summed survey time.
>
> res<- mutate(ddply(dat1,.(SPEC_CODE,Location_name),function(x) colSums(x[,-c(1:3,6)])),SumAIbySumST=Survey_Time/AI,Std.SumAIbySumST=10*SumAIbySumST)
> Thanks again for your input and help.
>
> Bruce
>


-- 
Bruce W. Miller, Ph.D.
Conservation Ecologist
Neotropical Bat Projects


office details
11384 Alpine Road
Stanwood, Mi. 49346
Phone (231) 679-6059