MEAT PRODUCTION

names(global_meat_production) <- make.names(names(global_meat_production))
### Summarize Data Set: global_meat_production
summary(global_meat_production)

production.df <- global_meat_production
production.ts <- ts(production.df, start=1961,end=2020, freq=12)

#global_meat_production
head(production.ts)
#production.ts
production.ts=as.data.frame( production.ts)

meat.ts <- ts( production.ts$Meat_total_.Production_.tonnes, start=1961, freq=12)
head(meat.ts)

summary(meat.ts)

plot(meat.ts, main='Meat Production Time Series', ylab='Meat Production')

acf(meat.ts)

ggmonthplot(meat.ts)
ggseasonplot(meat.ts)
ggsubseriesplot(meat.ts)

meatLog <- log(meat.ts)

plot(meatLog, main='Meat Production Time Series', ylab='Log of Meat Production')

acf(meatLog)

print(lambda <- BoxCox.lambda(meat.ts))  # find the optimal value for parameter lambda

plot.ts(BoxCox(meat.ts, lambda = lambda))

plot(meatLog, main='Meat Production Time Series', ylab='Log of Meat Production')

plot(BoxCox(meat.ts, lambda = lambda), main='Meat Production Time Series', ylab='BoxCox with lambda = -0.03151884')

acf(meatLog)

acf(BoxCox(meat.ts, lambda = lambda))

summary(fit <- lm(meatLog~time(meatLog)))

tsplot(meatLog, ylab="Meat Production", col=4, lwd=2)

tsplot(resid(fit), main="detrended")

acf(resid(fit))

summary(fit <- lm(meatLog ~ poly(time(meatLog), 2, raw=TRUE)))

tsplot(meatLog, ylab="Meat Production", col=4, lwd=2)

tsplot(resid(fit), main="detrended")

acf(resid(fit))

var(meatLog)

tsplot(meatLog ,main="Meat Production")

tsplot(diff(meatLog), main="first difference")

var(diff(meatLog))

acf(meatLog)

acf(resid(fit))

acf(diff(meatLog))

meatLog.stl <- stl(meatLog, "periodic")

plot(meatLog.stl)  # original series

meatLog.sa <- seasadj(meatLog.stl)

plot(meatLog)  # original series

plot(meatLog.sa)  # seasonal adjusted

acf(meatLog.sa)

seasonplot( meatLog.sa, 12, col=rainbow(12), year.labels=TRUE, main="Seasonal plot: Meat production") # seasonal frequency set as 12 for monthly data.

meatLog.sa <- seasadj(meatLog.stl)

summary(fit <- lm( meatLog.sa ~ poly(time( meatLog.sa), 2, raw=TRUE)))

tsplot(meatLog.sa, ylab="Meat Production", col=4, lwd=2)

tsplot(resid(fit), main="Residuals")

acf(resid(fit))

remainder.ts <-meatLog.stl$time.series[,'remainder']

acf(remainder.ts)

pacf(remainder.ts)

meat.arima = arima(remainder.ts, order = c(1,0,0), include.mean=F)

meat.arima

acf(meat.arima$resid[-1])

pacf(meat.arima$resid[-1])

var(meatLog)

var(diff(meatLog))

nsdiffs(meatLog)

ndiffs(meatLog)

plot(diff(meatLog,12))

acf(diff(meatLog,12))

var(diff(meatLog,12))

plot(diff(diff(meatLog,12)))

acf(diff(diff(meatLog,12)))

var(diff(diff(meatLog,12)))

residuals <- meat.arima$resid[-1]

plot(residuals)

plot.ts(residuals)

acf(residuals)

Correlation Test: Meat_total_.Production_.tonnes, Year

with(global_meat_production, cor.test(Meat_total_.Production_.tonnes, Year, 
  alternative="two.sided", method="pearson"))

Correlation Test: Meat_total_.Production_.tonnes, Year

with(global_meat_production, cor.test(Meat_total_.Production_.tonnes, Year, 
  alternative="two.sided", method="pearson"))

Normality Test: ~Meat_total_.Production_.tonnes

normalityTest(~Meat_total_.Production_.tonnes, test="ad.test", 
  data=global_meat_production)

Single-Sample t-Test: Meat_total_.Production_.tonnes

with(global_meat_production, (t.test(Meat_total_.Production_.tonnes, 
  alternative='two.sided', mu=0.0, conf.level=.95)))

Histogram: Meat_total_.Production_.tonnes

with(global_meat_production, Hist(Meat_total_.Production_.tonnes, 
  scale="frequency", breaks="Sturges", col="darkgray"))

Boxplot: ~ Year

Boxplot( ~ Year, data=global_meat_production, id=list(method="y"))

Scatterplot: Meat_total_.Production_.tonnes~Year

scatterplot(Meat_total_.Production_.tonnes~Year, regLine=FALSE, 
  smooth=FALSE, boxplots=FALSE, data=global_meat_production)