rm(list=ls())
setwd("C:/Users/wdbeavis/Desktop/PBEA/QG/Module 7/")
options(max.print=1000000)

#Next load in the data set from a spreadsheet (excel.csv)
# and 
ds <-read.csv("QG_Mod7_ALA7.1_ds.csv",header=T)
summary(ds)

ds$sample<-factor(ds$sample)
ds$field<-factor(ds$field)
ds$cooperator<-factor(ds$cooperator)
summary(ds)


########################################################################################
# next consider conducting EDA 
# do this so that you have estimates of experimental 
# variability that can be used to make decisions about 
# the quality of the data 

# the model for ANOVA is y = mean + coop + coop(field) + g ==sample + residual
# in this case, the model represents an RCBD
########################################################################################

attach(ds)
hist(ds$yield,col="beige",main="Frequency of yield ",
     xlab="yield")
boxplot(ds$yield~ds$sample,main="Boxplot by tc entry",
        xlab="Sample",ylab="Yield")
boxplot(ds$yield~ds$cooperator,main="Boxplot by Cooperator",
        xlab="Cooperator",ylab="Yield")
anova(lm(yield~cooperator+field/cooperator + sample))
detach(ds)