#Clear everything in the R console
rm(list=ls(all = TRUE))

#Install needed packages
install.packages("tseries")
install.packages("astsa")
install.packages("lmtest")
install.packages("moments")
install.packages("quantmod")
install.packages("plyr")
install.packages("Matrix")
#--- Please select a CRAN mirror for use in this session ---
library(quantmod)
library(ggplot2)
library(plyr)
library(moments)
#==============================================================#
#Basic operations
a <- 3
b <- 2 / 3
print(class(a))
print(a+b)
print(a*b)
print(a**2)
print(a^2)
print(paste0("a square is:", a**2))
print(as.integer(3.3))
print(class(as.integer(3.3)))
print(class(as.double(3)))

#Strings
Str <- c('My', 'name', 'is')
class(Str)
print(Str[1])
Str2 <- c(Str, 'Name')
Str2 <- append(Str, "Name")
Str3 <- append(Str2, "first", after = 1)

#Doubles
Dou_1 <- c(1,2,3)
Dou_2 <- c(1:10)
Dou_3 <- 1:5
Even <- seq(from = 2, to = 30, by = 2)

#Print Even divided by 2, multipied by 2, the first element of Even, the last element of Even
#
#
#
#
print(Even / 2)
print(Even*2)
print(Even[1]) #first element
print(Even[2:4]) #2nd, 3rd and 4th element of Even
print(Even[-1]) #all but first element
print(Even[-2]) #all but 2nd element
print(Even[-c(1,3)]) #all but first and 3rd elements
print(Even[length(Even)])

#Lists
List <- list(1:10,2:5)
print(List[1]) #1:10
print(List[[1]][1]) #1
c(List[1],List[2]) #or append(List[1], List[2])

#Sort the list in ascending and descending orders, using sort
sort(unlist(List))
sort(unlist(List), decreasing = TRUE)
#sort sorts the vector, order returns the indices of the vector in a sorted order
sort(c(3, 1, 2, 5, 4)) #1 2 3 4 5
order(c(3, 1, 2, 5, 4)) #2 3 1 5 4

#Arrays
Arr <- array(1:12, dim=c(3,4))
print(class(Arr)) #matrix
Arr_Ones <- array(1:1, dim=c(3,4))
print( Arr * Arr_Ones)
#Print the first row / column on Arr, the first 2 columns, the 1st and 3rd columns, the dimensions of Arr, add a column to Arr, sort Arr by increasing / decreasing order
#
#
#
#
print(Arr[1,]) #First row of Arr
print(Arr[,1]) #First column of Arr
print(Arr[,-1]) #Arr minus 1st column
print(Arr[,1:2]) #First two coolumns of Arr
print(Arr[,c(1,3)]) #First and third column of Arr
print(dim(Arr))
print(Arr[Arr>3]) #All elements of Arr >3
print(cbind(Arr, c(1,1,1))) #Add column
print(rbind(Arr, c(1,1,1,1))) #Add row
print(sort(Arr))
print(sort(Arr), decreasing = TRUE)
print(Arr[sort(-Arr[,3]),]) #sort Arr by 3rd column decreasing order

B <- matrix(c(2,4,6,1,3,5), nrow = 3, ncol = 2) #will fill by columns
C <- matrix(c(2,4,6,1,3,5), nrow = 2, ncol = 3, byrow = TRUE)
t(B) #transpose
c(B) #to get all the elements of the matrix B
#Get the rank of a matrix
library(Matrix)
rankMatrix(B)[1] #or qr(B)$rank
qr(array(0:0, dim = c(2,2)))$rank

#Calculate B squared, B multiplied by 2 and C divided by 2, Multiply the two matrices B and C
#
#
#
#
print(B^2)
print(2*B)
print(C / 2)
print(B %*% C)
print(C %*% B)

#Create identity matrix
Id <- diag(nrow = 2, ncol = 2)
#Calculate determinant
det(Id)
#Calculate inverse
qr.solve(2*Id)
#Sanity check: qr.solve(2*Id)*(2*Id)
#==============================================================#
#Generate random numbers and distributions

#Generate random numbers
runif(1) #between 0 and 1
runif(10) #generate 10 random numbers
runif(3, min=0, max=100)

#Generate random numbers following the Normal distribution
rnorm(10,0,1) #(n,mean,var)
#Generate 10000 rdm numbers following N(0,1) and plot their density 
#
#
#
#
plot(density(rnorm(10000,0,1)))

#==============================================================#
#Functions
Multi <- function(a,b){
	return(a*b)
}
nb = Multi(3,4)
nb = Multi(1,2)

#if statement
if (nb>10){
	print("Number greater than 10")
} else{
	print("Number smaller than 10")
}

#For loop
summ = 0
for (i in 2:10) {
	summ = summ + i
}
print(summ)

#Exercise:
#Create a function to check if a number is prime or not
#
#
#
#

isPrime <- function(n){
	flag = 0
	if (n>1) {
		flag = 1
		for (i in 2:(n-1)) {
			if ((n %% i) == 0) {
				flag = 0
				break
			}
			else{
				flag = 1
			}
		}
	}
	if (n ==2){
		flag = 1
	}
	if (flag == 1){
		print(paste(n, "is a prime number"))
	}
	else{
		print(paste(n, "is not a prime number"))
	}
}

isPrime(2)
isPrime(10)
isPrime(7)

#==============================================================#
#Import financial data from Yahoo Finance / Google
getSymbols('MSFT',src='yahoo')
MSFT #Gives Open, High, Low, Close prices, Volume and Adjusted Close price
MSFT[,1] #gives the Open prices

#Select time frame for Download
#auto.assign=FALSE: the data will be returned from the call, and will require the user to assign the results himself 
MSFT <- getSymbols('MSFT',src='yahoo', from = "2018-01-01", to = Sys.Date(), auto.assign=FALSE)
SP <- getSymbols('^GSPC',src='yahoo', from = "2018-01-01", to = Sys.Date(), auto.assign=FALSE)

#Get monthly data (daily, weekly or monthly)
MSFT <- getSymbols('MSFT',src='yahoo', from = "2018-01-01", to = Sys.Date(), periodicity = "monthly")

#==============================================================#
#Dealing with data
#Nb of rows and columns
nrow(MSFT)
ncol(MSFT)
#Visualise first few rows
head(MSFT)
tail(MSFT)
print(index(MSFT))

plot.ts(MSFT[,1])
plot.ts(MSFT$MSFT.Open)
#Rename columns
colnames(SP)[1] <- "SP.Open"
#Delete column with Adjusted prices
MSFT <- MSFT[,-6] #or MSFT$MSFT.Adjusted <- NULL
#Get number of missing values per column
colSums(is.na(MSFT))

#Print summary statistics
library(moments)
summary(MSFT)
mean(MSFT[,1], na.rm = TRUE)
sd(MSFT[,1], na.rm = TRUE) #standard deviation
skewness(MSFT, na.rm = TRUE)
kurtosis(MSFT, na.rm = TRUE)

#Delete missing data: 
data <- na.omit(MSFT)
#backward fill - last observation carried forward
data2 <- na.locf(MSFT)
#forward fill - Forward observation carried backwards
data3 <- na.locf(MSFT, fromlast = TRUE)


#Plot the Open and Closing Prices in the same graph
#Add legend, axis labels and title
ggplot(MSFT, aes(index(MSFT), y = value, color = Variable)) +
geom_line(aes(y = MSFT.Open, col = "Open prices")) +
geom_line(aes(y = MSFT.Close, col = "Close prices")) +
scale_color_manual(values = c("blue", "red")) +
labs(title = "MSFT Prices", x = "Date", y = "Prices")

#Plot the density of the Closing prices for MSFT
#Toy example: generate 1000 rdm numbers following N(0,1) and plot their density 
plot(density(rnorm(1000,0,1)))
#With financial data
plot(density(MSFT[,1]))

#==============================================================#
#How to merge two dataframes by date, and put NA values when missing data in one dataframe
library(zoo)
print(class(MSFT))
print(class(SP))
#Already zoo data, can merge directly, otherwise need to convert to zoo first, using
# MSFT <- read.zoo(MSFT)
mergData <- merge(MSFT, SP)
#merged Data by index, will put NA when to corresponding values
#Convert to DataFrame
Data <- as.data.frame(mergData)
colnames(Data)

#Scatter plot between SP and MSFT Closing Prices
#ann = FALSE to remove the labels
ggplot(Data, aes(x = Data$MSFT.Close, y = Data$GSPC.Close), ann = FALSE) + geom_point()

#Save dataframe
write.csv(Data, "/Users/Chloe/Downloads/Data.csv")
#Load data
dataset <- read.csv("/Users/Chloe/Downloads/Data.csv", header = TRUE)

#==============================================================#
#Download Microsoft and SP 500 data between 1/5/2019 and 30/9/2019
MSFT <- getSymbols('MSFT',src='yahoo', from = "2019-05-01", to ="2019-09-30", auto.assign=FALSE)
SP <- getSymbols('^GSPC',src='yahoo', from = "2019-05-01", to = "2019-09-30", auto.assign=FALSE)

#Calculate log Returns of MSFT Closing prices and plot them
#Toy example
Ex <- c(3:8)
Diff_Ex <- diff(Ex,1) #first difference of elements
diff(Ex, 2) #second difference of elements
#
#
#
#
log_Rt <- diff(log(MSFT$MSFT.Close),1)
head(log_Rt)
plot(index(log_Rt), log_Rt, type = "l")

#Add log_Rt to the mergData dataset and rename the column 'log_returns'
#
#
#
#
mergBis <- merge(MSFT, log_Rt)
colnames(mergBis)[ncol(mergBis)] <- "log_returns"
head(mergBis)

#Calculate lag of log_Rt
#
#
#
# 
head(lag(log_Rt,1))