| tuna {bayesm} | R Documentation |
Volume of canned tuna sales as well as a measure of display activity, log price and log wholesale price. Weekly data aggregated to the chain level. This data is extracted from the Dominick's Finer Foods database maintained by the University of Chicago http://http://research.chicagogsb.edu/marketing/databases/dominicks/dataset.aspx. Brands are seven of the top 10 UPCs in the canned tuna product category.
data(tuna)
A data frame with 338 observations on the following 30 variables.
WEEKMOVE1MOVE2MOVE3MOVE4MOVE5MOVE6MOVE7NSALE1NSALE2NSALE3NSALE4NSALE5NSALE6NSALE7LPRICE1LPRICE2LPRICE3LPRICE4LPRICE5LPRICE6LPRICE7LWHPRIC1LWHPRIC2LWHPRIC3LWHPRIC4LWHPRIC5LWHPRIC6LWHPRIC7FULLCUSTChevalier, A. Judith, Anil K. Kashyap and Peter E. Rossi (2003), "Why Don't Prices Rise During Periods of Peak Demand? Evidence from Scanner Data," The American Economic Review , 93(1), 15-37.
Chapter 7, Bayesian Statistics and Marketing by Rossi et al.
http://faculty.chicagogsb.edu/peter.rossi/research/bsm.html
data(tuna)
cat(" Quantiles of sales",fill=TRUE)
mat=apply(as.matrix(tuna[,2:5]),2,quantile)
print(mat)
##
## example of processing for use with rivGibbs
##
if(0)
{
data(tuna)
t = dim(tuna)[1]
customers = tuna[,30]
sales = tuna[,2:8]
lnprice = tuna[,16:22]
lnwhPrice= tuna[,23:29]
share=sales/mean(customers)
shareout=as.vector(1-rowSums(share))
lnprob=log(share/shareout)
# create w matrix
I1=as.matrix(rep(1, t))
I0=as.matrix(rep(0, t))
intercept=rep(I1, 4)
brand1=rbind(I1, I0, I0, I0)
brand2=rbind(I0, I1, I0, I0)
brand3=rbind(I0, I0, I1, I0)
w=cbind(intercept, brand1, brand2, brand3)
## choose brand 1 to 4
y=as.vector(as.matrix(lnprob[,1:4]))
X=as.vector(as.matrix(lnprice[,1:4]))
lnwhPrice=as.vector(as.matrix (lnwhPrice[1:4]))
z=cbind(w, lnwhPrice)
Data=list(z=z, w=w, x=X, y=y)
Mcmc=list(R=R, keep=1)
set.seed(66)
out=rivGibbs(Data=Data,Mcmc=Mcmc)
cat(" betadraws ",fill=TRUE)
summary(out$betadraw)
if(0){
## plotting examples
plot(out$betadraw)
}
}