[R] Conditional Weighted Average (ddply or any other function)

John Kane jrkrideau at inbox.com
Fri Mar 1 21:38:29 CET 2013


> For Sectors the results are correct, but for some (Country, Industry,
> FISCALYEAR) combinations or (Country,FISCALYEAR)combinations > the result don't match the spreadsheet (Excel) computation

Well the normal assumption here is to assume that Excel is wrong. I am not joking. Spreadsheets are notoriously error-prone, especially Excel, from problems with basic algorithms to 'simple' problems such as typos or accidentally specifying the wrong column of data. 

See http://www.r-bloggers.com/spreadsheet-errors/ [http://www.r-bloggers.com/spreadsheet-errors/] for one nasty example or from a recent thread on Excel and R 
"The idea that the Excel solver "has a good reputation for being fast and accurate" does not withstand an examination of the Excel solver's ability to solve the StRD nls test problems. Solver's ability is abysmal. 13 of 27 "answers" have zero accurate digits, and three more have fewer than two accurate digits -- and this is after tuning the solver to get a good answer. 
...
Excel solver does have the virture that it will always produce an answer, albeit one with zero accurate digits."
Bruce McCullough 

On the other hand there can be any number of reasons why R and Excel are not giving the same results assuming Excel may be correct. 

One thing that come to mind is that R and Excel may be handling missing data in a different manner if you have any missing data. 

Otherwise you may need to break out some data were the combination a)works with some combinations and does not with others and let some experts have a look at it. 

To my untutored eye you look like you're okay but I am not an expert by any means.

Sorry not to be of more help

John Kane
Kingston ON Canada

> -----Original Message-----
> From: anandpunit at gmail.com
> Sent: Fri, 1 Mar 2013 15:05:19 -0500
> To: jrkrideau at inbox.com
> Subject: Re: [R] Conditional Weighted Average (ddply or any other
> function)
> 
> For Sectors the results are correct, but for some (Country, Industry,
> FISCALYEAR) combinations or (Country, FISCALYEAR) combinations the
> result don't match the spreadsheet (Excel) computation, so verifying
> from experts, whether I am using ddply correctly with the right
> intention?
> 
> 
> On Fri, Mar 1, 2013 at 2:38 PM, John Kane <jrkrideau at inbox.com> wrote:
>> Is there any chance you meant
>> ddply (dat1 , .(Sector, FISCALYEAR), summarise, WROE=wavg(ROE, MKT))
>> ??
>> 
>> It gives a result but I have no idea if it makes sense.
>> 
>> John Kane
>> Kingston ON Canada
>> 
>> 
>>> -----Original Message-----
>>> From: anandpunit at gmail.com
>>> Sent: Fri, 1 Mar 2013 14:13:37 -0500
>>> To: r-help at r-project.org
>>> Subject: Re: [R] Conditional Weighted Average (ddply or any other
>>> function)
>>> 
>>> Hi John,
>>> 
>>> The sample size is huge involving 10,000 + firms. I have put a
>>> representative sample using dput ( Name, ticker and country have been
>>> changed so that firms cannot be identified due to proprietary data
>>> set, also EPS is not required and removed from the dataset)
>>> 
>>> structure(list(NAME = structure(c(8L, 8L, 8L, 8L, 8L, 8L, 7L,
>>> 7L, 7L, 7L, 7L, 7L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L,
>>> 6L, 6L, 4L, 4L, 4L, 4L, 4L, 4L, 3L, 3L, 3L, 3L, 3L, 3L, 1L, 1L,
>>> 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("CCC", "CTAX",
>>> "INN", "NOB", "SH", "SZ", "WASH", "WILLSON"), class = "factor"),
>>> Ticker = structure(c(7L, 7L, 7L, 7L, 7L, 7L, 8L, 8L, 8L,
>>> 8L, 8L, 8L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L,
>>> 6L, 4L, 4L, 4L, 4L, 4L, 4L, 3L, 3L, 3L, 3L, 3L, 3L, 1L, 1L,
>>> 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("CC13",
>>> "CT56", "INN12", "NB12", "SH12", "SZ12", "W12", "W15"), class =
>>> "factor"),
>>> Industry = structure(c(3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L,
>>> 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L,
>>> 4L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 5L, 5L,
>>> 5L, 5L, 5L, 5L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("Commercial &
>>> Professional Serv",
>>> "Energy", "Media", "Retail", "Transportation"), class = "factor"),
>>> Sector = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
>>> 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
>>> 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L,
>>> 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Consumer
>>> Discretionary",
>>> "Energy", "Industrials"), class = "factor"), Country =
>>> structure(c(4L,
>>> 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 3L, 3L, 3L, 3L,
>>> 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 1L, 1L, 1L,
>>> 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
>>> 2L, 2L, 2L), .Label = c("Brazil", "China", "India", "UK"), class =
>>> "factor"),
>>> FISCALYEAR = structure(c(3L, 2L, 1L, 4L, 5L, 6L, 3L, 2L,
>>> 1L, 4L, 5L, 6L, 3L, 2L, 1L, 4L, 5L, 6L, 3L, 3L, 2L, 1L, 4L,
>>> 5L, 6L, 3L, 2L, 1L, 4L, 5L, 6L, 3L, 2L, 1L, 4L, 5L, 6L, 3L,
>>> 2L, 1L, 4L, 5L, 6L, 3L, 2L, 1L, 4L, 5L, 6L), .Label = c("FY-1",
>>> "FY-2", "FY-3", "FY0", "FY1", "FY2"), class = "factor"),
>>> ROE = c(0.026, 0.0656, 0.1621, 0.1885, 0.1968, 0.2126, 0.0207,
>>> 0.0319, 0.0963, 0.0431, 0.066, 0.066, 0.0707, 0.0797, 0.0781,
>>> 0.078, 0.098, 0.126, 0.0352, 0.2625, 0.3714, 0.2929, 0.3133,
>>> 0.2509, 0.2398, 0.2779, 0.1109, 0.0509, 0.069, 0.1017, 0.1298,
>>> 0.5842, 0.3953, 0.4429, 0.3616, 0.26, 0.2, 0.4472, 0.2912,
>>> 0.21, 0.2849, 0.3553, 0.4347, 0.3289, 0.3846, 0.2643, 0.0458,
>>> 0.1265, 0.28), MKT = c(2919236084, 836858582, 2015182617,
>>> 3399344971, 4324821777, 4324821777, 7619453125, 3579844727,
>>> 4132238281, 3712239990, 2879757813, 2879757813, 1525237793,
>>> 700357605, 1814942993, 1858225342, 1242890503, 1242890503,
>>> 1879700000, 557093400, 224900300, 1634700000, 1443200000,
>>> 3582664735, 3582664735, 5830366211, 10660833984, 9024061523,
>>> 7628660645, 9154108398, 9154108398, 7064532227, 1804380005,
>>> 6331067871, 10445639648, 9153587891, 9153587891, 6231200000,
>>> 4.078e+09, 10107500000, 12460300000, 17800051556, 17800051556,
>>> 513478700, 260993500, 882575400, 1.151e+09, 855938413, 855938413
>>> )), .Names = c("NAME", "Ticker", "Industry", "Sector", "Country",
>>> "FISCALYEAR", "ROE", "MKT"), class = "data.frame", row.names = c(NA,
>>> -49L))
>>> 
>>> Thanks,
>>> Punit
>>> 
>>>> On Fri, Mar 1, 2013 at 12:51 PM, John Kane <jrkrideau at inbox.com>
>>>> wrote:
>>>> See below
>>>> 
>>>> 
>>>>> -----Original Message-----
>>>>> From: anandpunit at gmail.com
>>>>> Sent: Fri, 1 Mar 2013 12:36:53 -0500
>>>>> To: jrkrideau at inbox.com
>>>>> Subject: Re: [R] Conditional Weighted Average (ddply or any other
>>>>> function)
>>>>> 
>>>>> Hi John,
>>>>> 
>>>>> I was using symbols, Column ROE, EPS, MKTCAP are numeric, Name,
>>>>> Ticker, Sector, Country, FISCALYEAR or Year are character strings.
>>>>> 
>>>>> and column "Year" is referring to "FISCALYEAR"
>>>>> 
>>>> Definitely a no-no in R-help. :) We really need some
>>>> representative
>>>> sample data to play with. See
>>>> https://github.com/hadley/devtools/wiki/Reproducibility [https://github.com/hadley/devtools/wiki/Reproducibility] for some
>>>> general
>>>> pointers on how to compose a good question. The fact that you
>>>> included
>>>> the code you are using was excellent but without some data it is
>>>> rather
>>>> useless.
>>>> 
>>>> The easiest way to supply data is to use the dput() function.
>>>> Example
>>>> with your file named "testfile":
>>>> dput(testfile)
>>>> Then copy the output and paste into your email. This is what I did
>>>> with
>>>> your data that I pasted into my email . I added the dat1 <- to it.
>>>> 
>>>> For large data sets, you can just supply a representative sample.
>>>> Usually, dput(head(testfile, 100)) will be sufficient.
>>>> 
>>>> I hope this is of some help.
>>>> 
>>>> 
>>>>> 
>>>>> On Fri, Mar 1, 2013 at 12:31 PM, John Kane <jrkrideau at inbox.com>
>>>>> wrote:
>>>>>> It is not at all clear what you are doing. You state that the data
>>>>>> set
>>>>>> you are using is what I have called dat1 : see dput form below.
>>>>>> 
>>>>>> As far as I can see there is no numerical value in there.
>>>>>> 
>>>>>> ##===========data set in dput form================#
>>>>>> dat1 <- structure(list(Name = c("N1", "N1", "N1", "N1", "N1",
>>>>>> "N1",
>>>>>> "N1",
>>>>>> "N2", "N2", "N2", "N2", "N2", "N2", "N2"), Ticker = c("T1",
>>>>>> "T1",
>>>>>> "T1", "T1", "T1", "T1", "T1", "T2", "T2", "T2", "T2", "T2",
>>>>>> "T2",
>>>>>> "T2"), Sector = c("S1", "S1", "S1", "S1", "S1", "S1", "S1",
>>>>>> "S2",
>>>>>> "S2", "S2", "S2", "S2", "S2", "S2"), Industry = c("I1",
>>>>>> "I1",
>>>>>> "I1", "I1", "I1", "I1", "I1", "I2", "I2", "I2", "I2", "I2", "I2",
>>>>>> "I2"), Country = c("C1", "C1", "C1", "C1", "C1", "C1",
>>>>>> "C1",
>>>>>> "C2", "C2", "C2", "C2", "C2", "C2", "C2"), Year =
>>>>>> c("FY-4",
>>>>>> "FY-3",
>>>>>> "FY-2", "FY-1", "FY0", "FY1", "FY2", "FY-4", "FY-3",
>>>>>> "FY-2",
>>>>>> "FY-2", "FY0", "FY2", "FY2"), ROE = c("ROE11", "ROE12",
>>>>>> "ROE13",
>>>>>> "ROE14", "ROE15", "ROE16", "ROE17", "ROE21", "ROE22",
>>>>>> "ROE23",
>>>>>> "ROE24", "ROE25", "ROE26", "ROE27"), EPS = c("EPS11",
>>>>>> "EPS12",
>>>>>> "EPS13", "EPS14", "EPS15", "EPS16", "EPS17", "EPS21",
>>>>>> "EPS22",
>>>>>> "EPS23", "EPS24", "EPS25", "EPS26", "EPS27"), MKTCAP =
>>>>>> c("MKT11",
>>>>>> "MKT12", "MKT13", "MKT14", "MKT15", "MKT16", "MKT17",
>>>>>> "MKT21",
>>>>>> "MKT22", "MKT23", "MKT24", "MKT25", "MKT26", "MKT27")),
>>>>>> .Names
>>>>>> = c("Name",
>>>>>> "Ticker", "Sector", "Industry", "Country", "Year", "ROE",
>>>>>> "EPS",
>>>>>> "MKTCAP"), class = "data.frame", row.names = c(NA, -14L))
>>>>>> ## =================end of dataset==================#
>>>>>> 
>>>>>> There is no FISCALYEAR variable that you specifed below
>>>>>> 
>>>>>>> ddply (dataread , .(Sector, FISCALYEAR), summarise, >
>>>>>>> WROE=wavg(ROE,
>>>>>>> MKTCAP)))
>>>>>> 
>>>>>> I think we need a bit more information.
>>>>>> 
>>>>>> John Kane
>>>>>> Kingston ON Canada
>>>>>> 
>>>>>> 
>>>>>>> -----Original Message-----
>>>>>>> From: anandpunit at gmail.com
>>>>>>> Sent: Fri, 1 Mar 2013 11:01:42 -0500
>>>>>>> To: r-help at r-project.org
>>>>>>> Subject: [R] Conditional Weighted Average (ddply or any other
>>>>>>> function)
>>>>>>> 
>>>>>>> Hello R community,
>>>>>>> 
>>>>>>> I am computing weighted average statistic by using ddply function:
>>>>>>> 
>>>>>>> My data set is:
>>>>>>> N1 T1 S1 I1 C1 FY-4 ROE11 EPS11 MKT11
>>>>>>> N1 T1 S1 I1 C1 FY-3 ROE12 EPS12 MKT12
>>>>>>> N1 T1 S1 I1 C1 FY-2 ROE13 EPS13 MKT13
>>>>>>> N1 T1 S1 I1 C1 FY-1 ROE14 EPS14 MKT14
>>>>>>> N1 T1 S1 I1 C1 FY0 ROE15 EPS15 MKT15
>>>>>>> N1 T1 S1 I1 C1 FY1 ROE16 EPS16 MKT16
>>>>>>> N1 T1 S1 I1 C1 FY2 ROE17 EPS17 MKT17
>>>>>>> N2 T2 S2 I2 C2 FY-4 ROE21 EPS21 MKT21
>>>>>>> N2 T2 S2 I2 C2 FY-3 ROE22 EPS22 MKT22
>>>>>>> N2 T2 S2 I2 C2 FY-2 ROE23 EPS23 MKT23
>>>>>>> N2 T2 S2 I2 C2 FY-2 ROE24 EPS24 MKT24
>>>>>>> N2 T2 S2 I2 C2 FY0 ROE25 EPS25 MKT25
>>>>>>> N2 T2 S2 I2 C2 FY2 ROE26 EPS26 MKT26
>>>>>>> N2 T2 S2 I2 C2 FY2 ROE27 EPS27 MKT27
>>>>>>> 
>>>>>>> with colnames:
>>>>>>> (Name,Ticker,Sector,Industry,Country,Year,ROE,EPS,MKTCAP)
>>>>>>> 
>>>>>>> I want to compute
>>>>>>> 1) Weighted ROE based on Sector and Fiscal Year.
>>>>>>> For firm N1 of Sector S1 and Fiscalyear FY-3 weight is
>>>>>>> MKT1 / SUM(MKT, where Sector = S1, Fiscalyear FY-3)
>>>>>>> 
>>>>>>> 2) Weighted ROE based on Country and Fiscal Year.
>>>>>>> For firm N1 of Country C1 and Fiscalyear FY-3 weight is
>>>>>>> MKT1 / SUM(MKT, where Country = C1, Fiscalyear FY-3)
>>>>>>> 
>>>>>>> 3) Weighted ROE based on Country, Sector and Fiscal Year.
>>>>>>> For firm N1 of Country C1, Sector S1 and Fiscalyear FY-3
>>>>>>> weight is MKT1 / SUM(MKT, where Country = C1, Sector = S1,
>>>>>>> Fiscalyear
>>>>>>> FY-3)
>>>>>>> 
>>>>>>> 4) Weighted ROE based on Country, Industry and Fiscal Year.
>>>>>>> For firm N1 of Country C1, Industry I1 and Fiscalyear FY-3
>>>>>>> weight is MKT1 / SUM(MKT, where Country = C1, Industry = I1,
>>>>>>> Fiscalyear
>>>>>>> FY-3)
>>>>>>> 
>>>>>>> 
>>>>>>> I tried using ddply function:
>>>>>>> ddply (dataread , .(Sector, FISCALYEAR), summarise, WROE=wavg(ROE,
>>>>>>> MKTCAP)))
>>>>>>> 
>>>>>>> where wavg <- function(x, wt) x %*% wt/sum(wt)
>>>>>>> but this doesn't give me the right answer.
>>>>>>> 
>>>>>>> I could try subseting the data into different sectors and compute
>>>>>>> the
>>>>>>> weighted average which doesn't look like an elegant solution and
>>>>>>> would
>>>>>>> defeat the purpose of ddply
>>>>>>> 
>>>>>>> I coudn't think of properly using melt and cast functions to solve
>>>>>>> this issue. Any help will be highly appreciated.
>>>>>>> 
>>>>>>> Thanks and Regards,
>>>>>>> Punit
>>>>>>> 
>>>>>>> ______________________________________________
>>>>>>> R-help at r-project.org mailing list
>>>>>>> https://stat.ethz.ch/mailman/listinfo/r-help [https://stat.ethz.ch/mailman/listinfo/r-help]
>>>>>>> PLEASE do read the posting guide
>>>>>>> http://www.R-project.org/posting-guide.html [http://www.R-project.org/posting-guide.html]
>>>>>>> and provide commented, minimal, self-contained, reproducible code.
>>> 
>>> ______________________________________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help [https://stat.ethz.ch/mailman/listinfo/r-help]
>>> PLEASE do read the posting guide
>>> http://www.R-project.org/posting-guide.html [http://www.R-project.org/posting-guide.html]
>>> and provide commented, minimal, self-contained, reproducible code.
>> 
>> ____________________________________________________________
>> GET FREE SMILEYS FOR YOUR IM & EMAIL - Learn more at
>> http://www.inbox.com/smileys [http://www.inbox.com/smileys]
>> Works with AIM®, MSN® Messenger, Yahoo!® Messenger, ICQ®, Google Talk™
>> and most webmails
>> 
>>

____________________________________________________________
FREE 3D MARINE AQUARIUM SCREENSAVER - Watch dolphins, sharks & orcas on your desktop!



More information about the R-help mailing list