ds=read.csv("micro_math.csv")
str(ds)
summary(ds)
xtabs(~boring,data=ds)
xtabs(~boring+sex,data=ds)
boxplot(boring~sex,data=ds)
hist(ds$boring)
hist(ds$boring,breaks=c(.5,1.5,2.5,3.5,4.5))

fisher.test(xtabs(~boring+sex,data=ds))
fisher.test(xtabs(~useless+sex,data=ds))
fisher.test(xtabs(~hard+sex,data=ds))
fisher.test(xtabs(~scary+sex,data=ds))

t.test(ds$boring[ds$sex==1],ds$boring[ds$sex==2])
t.test(ds$useless[ds$sex==1],ds$useless[ds$sex==2])
t.test(ds$hard[ds$sex==1],ds$hard[ds$sex==2])
t.test(ds$scary[ds$sex==1],ds$scary[ds$sex==2])

ds$sum=ds$boring+ds$useless+ds$hard+ds$scary

rowSums(ds[,1:4])
t.test(ds$sum[ds$sex==1],ds$sum[ds$sex==2])



---
pme = function(i){p=t.test(ds[ds$sex==1,i],ds[ds$sex==2,i])$p.value;
return(p)}
> pme(2)
[1] 0.161473
> sapply(1:4,pme)
[1] 0.6327317 0.1614730 0.8225917 0.1069281


colSds=function(i){return(sd(ds[,i]))}
sapply(1:4,colSds)
> pairs(ds[,1:4])
> cor(ds[,1:4])
           boring   useless      hard     scary
boring  1.0000000 0.3218986 0.4378138 0.6882459
useless 0.3218986 1.0000000 0.2310647 0.2775823
hard    0.4378138 0.2310647 1.0000000 0.5333577
scary   0.6882459 0.2775823 0.5333577 1.0000000

library("corrplot")
library("Hmisc")
Dr=rcorr(as.matrix(ds))
Dr$P
xtabs(~boring+scary,data=ds)

corrplot(Dr$r, type="upper", order="hclust", tl.col="black", tl.srt=45)

ds$sum=rowSums(ds[,1:4])
----

qbinom(.05/2,10^6,.5)
qbinom(1-.05/2,10^6,.5)

curve(qbinom(x,10^6,.5),.01,.99)

plot(dbinom(499000:501000,10^6,.5))
sum(dbinom(0:499000,10^6,.5))
sum(dbinom(499000:501000,10^6,.5))

---
df=read.csv("titanic.csv")
str(df)
summary(df)

mean(df$Survived[df$Sex=="female"])
mean(df$Survived[df$Sex=="male"])
mean(df$Survived[df$Sex=="female" & df$Pclass==1])
sum(df$Survived[df$Sex=="female" & df$Pclass==1])
length(df$Survived[df$Sex=="female" & df$Pclass==1])

mean(df$Survived[df$Sex=="male" & df$Pclass==1])

xtabs(Survived~Sex+Pclass,data=df)
xtabs(!Survived~Sex+Pclass,data=df)