I am trying to create a normalization value for a variable I am working with based on individual conference means and SDs. I found the conference means using the function:
confavg=aggregate(base$AVG, by=list(base$confName), FUN=mean)
And so after getting the means for the 31 conferences, I want to go back and for each individual player put these means in so I can easily calculate a normalization factor based on the conference mean.
I have tried to create large ifelse or if statements where confavg is the conference average.
ifelse((base$confName=="America East Conference"),confavg[1,2]->base$CAVG,0->base$CAVG)
but nothing works. Ideally I would want to take every player and say:
Normalization = (player average - conference average)/conference standard deviation
How should I go about doing that?
edit:
Here is some sample data:
AVG = c(.350,.400,.320,.220,.100,.250,.400,.450)
Conf = c("SEC","ACC","SEC","B12","P12","ACC","B12","P12")
Conf=as.factor(Conf)
sampleconfavg=aggregate(AVG, by=list(Conf), FUN=mean)
sampleconfsd=aggregate(AVG, by=list(Conf), FUN=sd)
So each player would have their average - the conference average / sd of conference
so for the first guy it would be:
(.350 - .335) / 0.0212132 = 0.7071069
but I am hoping to build a function that does it for all people in my dataset. Thank you!
edit2:
Alright the answer below is amazing but I am running into (hopefully) one last problem. I want to basically do this process to three variables like:
base3=do.call(rbind, by(base3, base3$confName, FUN=function(x) { x$ScaledAVG <- scale(x$AVG); x}))
base3=do.call(rbind, by(base3, base3$confName, FUN=function(x) { x$ScaledOBP <- scale(x$OBP); x}))
base3=do.call(rbind, by(base3, base3$confName, FUN=function(x) { x$ScaledK.AB <- scale(x$K.AB); x}))
Which works but then when I search the datafile like:
base3[((base3$ScaledAVG>2)&(base3$ScaledOBP>2)&(base3$ScaledK.AB<.20)),]
it resets the Scaled K.AB value and doesn't use it as part of the parameters of the search.
Here is an example to scale iris$Sepal.Length, within groups of iris$Species:
scaled.iris <- do.call(rbind,
by(iris, iris$Species,
FUN=function(x) { x$Scaled.Sepal.Length <- scale(x$Sepal.Length); x }
)
)
head(scaled.iris)
## Sepal.Length Sepal.Width Petal.Length Petal.Width Species Scaled.Sepal.Length
## setosa.1 5.1 3.5 1.4 0.2 setosa 0.26667447
## setosa.2 4.9 3.0 1.4 0.2 setosa -0.30071802
## setosa.3 4.7 3.2 1.3 0.2 setosa -0.86811050
## setosa.4 4.6 3.1 1.5 0.2 setosa -1.15180675
## setosa.5 5.0 3.6 1.4 0.2 setosa -0.01702177
## setosa.6 5.4 3.9 1.7 0.4 setosa 1.11776320
Edit:
Using your sample data (Conf and AVG only):
d <- data.frame(Conf, AVG)
dd <- do.call(rbind, by(d, d$Conf, FUN=function(x) { x$Scaled <- scale(x$AVG); x}))
# Remove generated row names
rownames(dd) <- NULL
dd
## Conf AVG Scaled
## 1 ACC 0.40 0.7071068
## 2 ACC 0.25 -0.7071068
## 3 B12 0.22 -0.7071068
## 4 B12 0.40 0.7071068
## 5 P12 0.10 -0.7071068
## 6 P12 0.45 0.7071068
## 7 SEC 0.35 0.7071068
## 8 SEC 0.32 -0.7071068