Updating to latest version of labs
This commit is contained in:
John Lam
Родитель
925369cce1
Коммит
bce5c59aad
Двоичный файл не отображается.
|
|
@ -6,11 +6,13 @@ library(rhandsontable)
|
|||
starting.investment = 100000
|
||||
start.date = "2017-01-01"
|
||||
|
||||
compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
||||
portfolio <- cbind(portfolio, dollars = portfolio$percentage * dollars)
|
||||
compute.portfolio.daily.market.value <- function(portfolio, dollars) {
|
||||
portfolio <- cbind(
|
||||
portfolio,
|
||||
dollars = portfolio$percentage * dollars)
|
||||
df <- NULL
|
||||
|
||||
compute.daily.book.value <- function(symbol, dollars) {
|
||||
compute.daily.market.value <- function(symbol, dollars) {
|
||||
|
||||
symbol.data <- getSymbols(
|
||||
symbol,
|
||||
|
|
@ -18,28 +20,25 @@ compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
|||
from = start.date)
|
||||
|
||||
shares <- dollars / as.numeric(first(Op(symbol.data)))
|
||||
book.value <- Ad(symbol.data) * shares
|
||||
market.value <- Ad(symbol.data) * shares
|
||||
|
||||
if (is.null(df)) {
|
||||
df <<- data.frame(book.value)
|
||||
df <<- data.frame(market.value)
|
||||
}
|
||||
else {
|
||||
df <<- cbind(df, data.frame(book.value))
|
||||
df <<- cbind(df, data.frame(market.value))
|
||||
}
|
||||
}
|
||||
|
||||
mapply(compute.daily.book.value, portfolio$symbols, portfolio$dollars)
|
||||
|
||||
# Compute totals for each day
|
||||
mapply(
|
||||
compute.daily.market.value,
|
||||
portfolio$symbols,
|
||||
portfolio$dollars)
|
||||
|
||||
return(cbind(df, data.frame(Total = rowSums(df))))
|
||||
}
|
||||
|
||||
convert.totals.dataframe.to.xts <- function(df) {
|
||||
return(xts(df$Total, as.POSIXct(rownames(df))))
|
||||
}
|
||||
|
||||
function(input, output) {
|
||||
shinyServer(function(input, output) {
|
||||
|
||||
v <- reactiveValues(portfolio = NULL, index = NULL)
|
||||
|
||||
|
|
@ -52,8 +51,10 @@ function(input, output) {
|
|||
|
||||
portfolio = hot_to_r(input$hot)
|
||||
|
||||
v$portfolio <- compute.portfolio.daily.book.value(portfolio, starting.investment)
|
||||
v$index <- compute.portfolio.daily.book.value(
|
||||
v$portfolio <- compute.portfolio.daily.market.value(
|
||||
portfolio,
|
||||
starting.investment)
|
||||
v$index <- compute.portfolio.daily.market.value(
|
||||
data.frame(
|
||||
symbols = c("QQQ"),
|
||||
percentage = c(1.0),
|
||||
|
|
@ -61,14 +62,21 @@ function(input, output) {
|
|||
),
|
||||
starting.investment
|
||||
)
|
||||
print(v$index)
|
||||
})
|
||||
|
||||
convert.totals.dataframe.to.xts <- function(df) {
|
||||
return(xts(df$Total, as.POSIXct(rownames(df))))
|
||||
}
|
||||
|
||||
output$chart <- renderHighchart({
|
||||
if (is.null(v$portfolio) || is.null(v$index)) return()
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(v$portfolio), name = "Portfolio") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(v$index), name = "QQQ") %>%
|
||||
hc_add_series(
|
||||
data = convert.totals.dataframe.to.xts(v$portfolio),
|
||||
name = "Portfolio") %>%
|
||||
hc_add_series(
|
||||
data = convert.totals.dataframe.to.xts(v$index),
|
||||
name = "QQQ") %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
})
|
||||
|
||||
|
|
@ -85,4 +93,4 @@ function(input, output) {
|
|||
rhandsontable(portfolio) %>%
|
||||
hot_table(highlightCol = TRUE, highlightRow = TRUE)
|
||||
})
|
||||
}
|
||||
})
|
||||
|
|
@ -4,103 +4,205 @@
|
|||
library(quantmod)
|
||||
|
||||
# We are comparing our performance against QQQ
|
||||
# which tracks the Nasdaq 100
|
||||
# which tracks the Nasdaq 100 index. Run the lines of code
|
||||
# below to get all the data from the beginning of the year.
|
||||
|
||||
start.date <- "2017-01-01"
|
||||
QQQ <- getSymbols("QQQ", auto.assign = FALSE, from = start.date)
|
||||
|
||||
# View this in variable explorer
|
||||
# Open table viewer
|
||||
# Export to Excel
|
||||
# View the contents of the data frame in variable explorer. Try:
|
||||
#
|
||||
# 1. Clicking on the chevron to the left of the QQQ variable to
|
||||
# drill down into its internal structure.
|
||||
#
|
||||
# 2. Clicking on the hourglass icon to the right of the QQQ
|
||||
# variable to open up the data in the Data Table viewer.
|
||||
#
|
||||
# 3. Clicking on the Excel icon to the right of the QQQ variable
|
||||
# to open the data frame in Excel.
|
||||
|
||||
# Now lets plot QQQ for the YTD using the HighCharter package
|
||||
# You should see a chart open up in the default browser, showing
|
||||
# a plot of the Open, High, Low, and Close prices for each date.
|
||||
# Note that the highcharter package is simply a wrapper around
|
||||
# the HighChart commercial library. See more at
|
||||
# https://highcharts.com.
|
||||
|
||||
library(highcharter)
|
||||
|
||||
# Notice the %>% operator. It's also known as the "pipe" operator
|
||||
# which lets you chain together multiple function calls into a
|
||||
# single statement. It's provided by the magrittr library.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ, type = "ohlc")
|
||||
|
||||
# Let's get another stock MSFT and plot that in a comparison chart
|
||||
# Next, let's grab data for another stock, MSFT, and create a chart
|
||||
# that compares the performance of MSFT vs. the QQQ S&P100 index.
|
||||
|
||||
MSFT <- getSymbols("MSFT", auto.assign = FALSE, from = start.date)
|
||||
|
||||
# Here, we generate a chart, showing both stocks head-to-head.
|
||||
# However, because the stock prices are at different scales, it's
|
||||
# hard to see a clear analysis of their performance. Also
|
||||
# notice how we've added the 538 (from 538.com) theme to the plot.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ, type = "ohlc") %>%
|
||||
hc_add_series(MSFT) %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# So we need to transform our data so that we look at the increase
|
||||
# in value of a portfolio invested equally between the two
|
||||
# series that we want to visualize. Before we can perform
|
||||
# calculations, we first need to take a look at the data structures.
|
||||
|
||||
# Let's examine the data again in QQQ. You can see that it contains
|
||||
# open, high, low, close, volume, and adjusted prices.
|
||||
# open, high, low, close, volume, and adjusted prices. The command
|
||||
# below first converts QQQ to a data frame so that we can visualize
|
||||
# it using our built in data grid viewer. Experiment with the grid
|
||||
# to play with the different features that it offers!
|
||||
|
||||
View(as.data.frame(QQQ))
|
||||
|
||||
# We really are only interested in the adjusted prices at close
|
||||
# for the purposes of computing daily returns, which we can
|
||||
# extract using the Ad() function
|
||||
# for the purposes of computing daily returns. We can extract the
|
||||
# adjusted prices column from the time series using the Ad()
|
||||
# function. Note that in R, the "." is just part of the variable
|
||||
# name; it has no semantic meaning.
|
||||
|
||||
QQQ.adjusted <- Ad(QQQ)
|
||||
|
||||
# Let's view the first few rows of the xts result set. We can see
|
||||
# that it contains both dates and the adjusted price for QQQ.
|
||||
# that it contains both dates and the adjusted price for QQQ. The
|
||||
# dates are part of the row labels (each row can have a label)
|
||||
# associated with it, and it's not a separate column. You can
|
||||
# extract the row labels using the rownames() function, which
|
||||
# we will see later in this lab.
|
||||
|
||||
View(as.data.frame(QQQ.adjusted))
|
||||
|
||||
# Now let's compute returns by buying a certain number of shares of the stock
|
||||
# Now let's compute returns by buying a certain number of shares
|
||||
# of the stock
|
||||
|
||||
starting.investment = 100000
|
||||
|
||||
# Compute the number of shares we can buy by dividing starting investment by opening price
|
||||
# Compute the number of shares we can buy by dividing starting
|
||||
# investment by opening price in the time period
|
||||
|
||||
MSFT.starting.shares = starting.investment / as.numeric(first(Op(MSFT)))
|
||||
QQQ.starting.shares = starting.investment / as.numeric(first(Op(QQQ)))
|
||||
|
||||
# Now compute the daily adjusted book value of our shares
|
||||
# This is a vector * scalar which yields another vector
|
||||
# Now compute the daily adjusted market value of our shares
|
||||
# This is a vector * scalar which yields another vector. Put
|
||||
# another way, we are multiplying each row in the adjusted closing
|
||||
# price by the # shares that we hold.
|
||||
|
||||
MSFT.daily.book.value = Ad(MSFT) * MSFT.starting.shares
|
||||
QQQ.daily.book.value = Ad(QQQ) * QQQ.starting.shares
|
||||
MSFT.daily.market.value = Ad(MSFT) * MSFT.starting.shares
|
||||
QQQ.daily.market.value = Ad(QQQ) * QQQ.starting.shares
|
||||
|
||||
# Now let's plot the value of our investments in QQQ vs. MSFT
|
||||
# on a single chart to see which one wins.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ.daily.book.value, name = "QQQ") %>%
|
||||
hc_add_series(MSFT.daily.book.value, name = "MSFT") %>%
|
||||
hc_add_series(QQQ.daily.market.value, name = "QQQ") %>%
|
||||
hc_add_series(MSFT.daily.market.value, name = "MSFT") %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# Now let's compute the book value based on the amount invested in a stock
|
||||
# Now let's compute the market value based on the amount invested
|
||||
# in an arbitrary stock. We'll define a new function,
|
||||
# compute.daily.market.value(), which takes a symbol, the amount
|
||||
# invested in it and returns a list that contains the market
|
||||
# value for each day in the investment range.
|
||||
|
||||
compute.daily.book.value <- function(symbol, dollars) {
|
||||
symbol.data <- getSymbols(symbol, auto.assign = FALSE, from = start.date)
|
||||
# There are some new notable R features in this example:
|
||||
# 1. Functions are defined as values
|
||||
# 2. The return statement returns the result of an expression. If
|
||||
# you don't use return, then the value of the last evaluated
|
||||
# expression will be returned from your function.
|
||||
|
||||
# When you run the function, you can either:
|
||||
# 1. Select all the code in the editor and press CTRL+ENTER to
|
||||
# send it in one chunk to the REPL.
|
||||
# 2. Send one line at a time by repeatedly pressing CTRL+ENTER
|
||||
# to see our multi-line support in the REPL. You'll see this
|
||||
# in the "+" characters running down the left margin to
|
||||
# indicate that you're continuing the current block of code
|
||||
# in the REPL.
|
||||
|
||||
compute.daily.market.value <- function(symbol, dollars) {
|
||||
symbol.data <- getSymbols(
|
||||
symbol,
|
||||
auto.assign = FALSE,
|
||||
from = start.date)
|
||||
shares <- dollars / first(Op(symbol.data))
|
||||
return(Ad(symbol.data) * shares)
|
||||
}
|
||||
|
||||
# Now, let's get the daily returns, expressed as a percentage
|
||||
# gain or loss, based on the adjusted prices
|
||||
# gain or loss, based on the adjusted prices. All we have to
|
||||
# do now is call the compute.daily.market.value() with the
|
||||
# name of the symbol, and the starting investment.
|
||||
|
||||
QQQ.daily.book.value <- compute.daily.book.value("QQQ", starting.investment)
|
||||
View(as.data.frame(QQQ.daily.book.value))
|
||||
QQQ.daily.market.value <- compute.daily.market.value(
|
||||
"QQQ",
|
||||
starting.investment)
|
||||
|
||||
# Let's test it on TSLA
|
||||
View(as.data.frame(QQQ.daily.market.value))
|
||||
|
||||
TSLA.daily.book.value <- compute.daily.book.value("TSLA", starting.investment)
|
||||
View(TSLA.daily.book.value)
|
||||
# Now let's do the same thing for an equal weight portfolio
|
||||
# of FANG (Facebook, Amazon, Netflix, Google). Let's define
|
||||
# a data frame that contains two columns: symbols and percentage
|
||||
# that represent the stock symbol names and their percentage
|
||||
# weight in the portfolio (these ideally add up to 1.0 -
|
||||
# enforcing that is an exercise left to the reader).
|
||||
|
||||
TSLA.cumulative.returns <- cumsum(TSLA.daily.returns * starting.investment)
|
||||
head(TSLA.cumulative.returns)
|
||||
|
||||
# Now let's do the same thing for an equal weight portfolio of FANG
|
||||
# You will notice that there is a peculiar named parameter in
|
||||
# the call to the data.frame() function to define the data frame.
|
||||
# stringsAsFactors is something you'll run into often in R -
|
||||
# by default, R interprets character strings as Factors. These
|
||||
# are roughly analogous to enumeration values in other languages.
|
||||
# You get better performance, but these values are not treated
|
||||
# as strings. This parameter forces the data.frame() function to
|
||||
# treat the list of strings assigned to the symbols column as
|
||||
# strings and not factors.
|
||||
|
||||
portfolio <- data.frame(
|
||||
symbols = c("FB", "AMZN", "NFLX", "GOOG"),
|
||||
percentage = c(0.25, 0.25, 0.25, 0.25),
|
||||
stringsAsFactors = FALSE)
|
||||
|
||||
# We need to adjust our daily book value function to compute against portfolio of stocks
|
||||
# Now we need to change our daily market value function to compute
|
||||
# against portfolio of stocks rather than a single stock.
|
||||
|
||||
compute.portfolio.daily.market.value <- function(portfolio, dollars) {
|
||||
|
||||
# The cbind() function binds two dataframes together by column.
|
||||
# In this case, we're appending a new column, dollars, to
|
||||
# the portfolio dataframe. That column contains the result
|
||||
# of muliplying each row in the percentage column against the
|
||||
# dollars scalar value.
|
||||
|
||||
# Note that there are no explicit loops here. In general, if
|
||||
# you find yourself writing a loop in R, you're probably
|
||||
# doing it wrong :)
|
||||
|
||||
portfolio <- cbind(
|
||||
portfolio,
|
||||
dollars = portfolio$percentage * dollars)
|
||||
|
||||
# We define a variable here in the middle of the body of
|
||||
# the compute.portfolio.daily.market.value() function. Because
|
||||
# R supports lexical scoping, this variable can be manipulated
|
||||
# from within the compute.daily.market.value() function that
|
||||
# is defined immediately below it.
|
||||
|
||||
compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
||||
portfolio <- cbind(portfolio, dollars = portfolio$percentage * dollars)
|
||||
df <- NULL
|
||||
|
||||
compute.daily.book.value <- function(symbol, dollars) {
|
||||
# This is a nested function. It can access variables in
|
||||
# the enclosing scope by using the <<- operator.
|
||||
|
||||
compute.daily.market.value <- function(symbol, dollars) {
|
||||
|
||||
symbol.data <- getSymbols(
|
||||
symbol,
|
||||
|
|
@ -108,46 +210,93 @@ compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
|||
from = start.date)
|
||||
|
||||
shares <- dollars / as.numeric(first(Op(symbol.data)))
|
||||
book.value <- Ad(symbol.data) * shares
|
||||
market.value <- Ad(symbol.data) * shares
|
||||
|
||||
# Here, we use the <<- operator to either:
|
||||
# 1. Assign a new data frame if one didn't exist already
|
||||
# 2. Append a new column to the existing data frame.
|
||||
|
||||
if (is.null(df)) {
|
||||
df <<- data.frame(book.value)
|
||||
df <<- data.frame(market.value)
|
||||
}
|
||||
else {
|
||||
df <<- cbind(df, data.frame(book.value))
|
||||
df <<- cbind(df, data.frame(market.value))
|
||||
}
|
||||
}
|
||||
|
||||
mapply(compute.daily.book.value, portfolio$symbols, portfolio$dollars)
|
||||
# This is how we iterate over all of the rows in the
|
||||
# portfolio data frame. We pass the symbols column to
|
||||
# the compute.daily.market.value() function by dereferencing
|
||||
# it using the $ operator. Similarly, we pass the dollars
|
||||
# column as well. When this function returns, the df
|
||||
# variable will now hold a new data frame containing the
|
||||
# daily market values of all of the stocks in the portfolio.
|
||||
|
||||
# Compute totals for each day
|
||||
mapply(
|
||||
compute.daily.market.value,
|
||||
portfolio$symbols,
|
||||
portfolio$dollars)
|
||||
|
||||
# Compute totals for each day using the rowSums function.
|
||||
# Here we create a new data frame that contains a column
|
||||
# called Total, which contains the sum of the daily
|
||||
# market values of each of the stock symbols.
|
||||
|
||||
df <- cbind(
|
||||
df,
|
||||
data.frame(Total = rowSums(df)))
|
||||
|
||||
df <- cbind(df, data.frame(Total = rowSums(df)))
|
||||
return(df)
|
||||
}
|
||||
|
||||
portfolio.daily.book.value <- compute.portfolio.daily.book.value(portfolio, starting.investment)
|
||||
# Here's a single function call that computes the daily market
|
||||
# value of our portfolio.
|
||||
|
||||
portfolio.daily.market.value <- compute.portfolio.daily.market.value(
|
||||
portfolio,
|
||||
starting.investment)
|
||||
|
||||
# Let's define another "portfolio", but this time it contains only
|
||||
# a single stock - QQQ which represents the NASDAQ 100 index.
|
||||
|
||||
index <- data.frame(
|
||||
symbols = c("QQQ"),
|
||||
percentage = c(1.0),
|
||||
stringsAsFactors = FALSE)
|
||||
|
||||
index.daily.book.value <- compute.portfolio.daily.book.value(index, starting.investment)
|
||||
index.daily.market.value <- compute.portfolio.daily.market.value(
|
||||
index,
|
||||
starting.investment)
|
||||
|
||||
# Now let's plot the two curves against each other
|
||||
# This loses the date labels - better to remove columns?
|
||||
# Our daily market value results are in a data frame.
|
||||
# We need to convert them to an xts object (time series object)
|
||||
# so that the dates display correctly in the plots. We didn't
|
||||
# need to do this in our previous single-stock calculations
|
||||
# because we modified an existing xts object in-place. In our
|
||||
# new portfolio base calculation, we are computing a new data
|
||||
# frame from scratch, hence the need to convert it to
|
||||
# an xts object using the labels from the data frame that
|
||||
# contain the identity of each trading day.
|
||||
|
||||
convert.totals.dataframe.to.xts <- function(df) {
|
||||
return(xts(df$Total, as.POSIXct(rownames(df))))
|
||||
return(
|
||||
xts(
|
||||
df$Total,
|
||||
as.POSIXct(rownames(df))))
|
||||
}
|
||||
|
||||
# Plot the comparison chart between our portfolio and QQQ.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(index.daily.book.value), name = "QQQ") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(portfolio.daily.book.value), name = "Portfolio") %>%
|
||||
hc_add_series(
|
||||
convert.totals.dataframe.to.xts(index.daily.market.value),
|
||||
name = "QQQ") %>%
|
||||
hc_add_series(
|
||||
convert.totals.dataframe.to.xts(portfolio.daily.market.value),
|
||||
name = "Portfolio") %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# Now the next step in this is to turn this analysis into an applicatoin
|
||||
# where you can pick the stocks to add to your asset allocation and their
|
||||
# proportions. The constraint is that they must add up to 100% in your allocation.
|
||||
# Hit the compute button to generate the results.
|
||||
# Now the next step in this is to turn this analysis into an
|
||||
# application where you can pick the stocks to add to your
|
||||
# asset allocation and their percentages. See the next
|
||||
# section in the lab manual for how to complete this part of the lab.
|
||||
|
|
|
|||
|
|
@ -0,0 +1,10 @@
|
|||
library(shiny)
|
||||
library(quantmod)
|
||||
library(highcharter)
|
||||
library(rhandsontable)
|
||||
|
||||
shinyServer(function(input, output) {
|
||||
observeEvent(input$go, {
|
||||
print("Hello, World!")
|
||||
})
|
||||
})
|
||||
|
|
@ -0,0 +1,14 @@
|
|||
library(shiny)
|
||||
library(highcharter)
|
||||
library(rhandsontable)
|
||||
|
||||
shinyUI(
|
||||
fluidPage(
|
||||
tags$link(rel = "stylesheet", type = "text/css", href = "https://bootswatch.com/paper/bootstrap.css"),
|
||||
fluidRow(
|
||||
column(width = 3, class = "panel",
|
||||
actionButton("go", label = "Say Hello")
|
||||
)
|
||||
)
|
||||
)
|
||||
)
|
||||
|
|
@ -21,16 +21,18 @@ QQQ <- getSymbols("QQQ", auto.assign = FALSE, from = start.date)
|
|||
# 3. Clicking on the Excel icon to the right of the QQQ variable
|
||||
# to open the data frame in Excel.
|
||||
|
||||
|
||||
# Now lets plot QQQ for the YTD using the HighCharter package
|
||||
# You should see a chart open up in the default browser, showing
|
||||
# a plot of the Open, High, Low, and Close prices for each date.
|
||||
# Note that the highcharter package is simply a wrapper around
|
||||
# the HighChart commercial library. See more at
|
||||
# https://highcharts.com.
|
||||
|
||||
library(highcharter)
|
||||
|
||||
# Notice the %>% operator. It's also known as the "pipe" operator
|
||||
# which lets you chain together multiple function calls into a
|
||||
# single statement.
|
||||
# single statement. It's provided by the magrittr library.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ, type = "ohlc")
|
||||
|
|
@ -42,85 +44,165 @@ MSFT <- getSymbols("MSFT", auto.assign = FALSE, from = start.date)
|
|||
|
||||
# Here, we generate a chart, showing both stocks head-to-head.
|
||||
# However, because the stock prices are at different scales, it's
|
||||
# hard to see a clear analysis of their performance.
|
||||
# hard to see a clear analysis of their performance. Also
|
||||
# notice how we've added the 538 (from 538.com) theme to the plot.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ, type = "ohlc") %>%
|
||||
hc_add_series(MSFT) %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# So we need to transform our data so that we look at the increase
|
||||
# in value of a portfolio invested equally between the two
|
||||
# series that we want to visualize. Before we can perform
|
||||
# calculations, we first need to take a look at the data structures.
|
||||
|
||||
# Let's examine the data again in QQQ. You can see that it contains
|
||||
# open, high, low, close, volume, and adjusted prices.
|
||||
# open, high, low, close, volume, and adjusted prices. The command
|
||||
# below first converts QQQ to a data frame so that we can visualize
|
||||
# it using our built in data grid viewer. Experiment with the grid
|
||||
# to play with the different features that it offers!
|
||||
|
||||
View(as.data.frame(QQQ))
|
||||
|
||||
# We really are only interested in the adjusted prices at close
|
||||
# for the purposes of computing daily returns, which we can
|
||||
# extract using the Ad() function
|
||||
# for the purposes of computing daily returns. We can extract the
|
||||
# adjusted prices column from the time series using the Ad()
|
||||
# function. Note that in R, the "." is just part of the variable
|
||||
# name; it has no semantic meaning.
|
||||
|
||||
QQQ.adjusted <- Ad(QQQ)
|
||||
|
||||
# Let's view the first few rows of the xts result set. We can see
|
||||
# that it contains both dates and the adjusted price for QQQ.
|
||||
# that it contains both dates and the adjusted price for QQQ. The
|
||||
# dates are part of the row labels (each row can have a label)
|
||||
# associated with it, and it's not a separate column. You can
|
||||
# extract the row labels using the rownames() function, which
|
||||
# we will see later in this lab.
|
||||
|
||||
View(as.data.frame(QQQ.adjusted))
|
||||
|
||||
# Now let's compute returns by buying a certain number of shares of the stock
|
||||
# Now let's compute returns by buying a certain number of shares
|
||||
# of the stock
|
||||
|
||||
starting.investment = 100000
|
||||
|
||||
# Compute the number of shares we can buy by dividing starting investment by opening price
|
||||
# Compute the number of shares we can buy by dividing starting
|
||||
# investment by opening price in the time period
|
||||
|
||||
MSFT.starting.shares = starting.investment / as.numeric(first(Op(MSFT)))
|
||||
QQQ.starting.shares = starting.investment / as.numeric(first(Op(QQQ)))
|
||||
|
||||
# Now compute the daily adjusted book value of our shares
|
||||
# This is a vector * scalar which yields another vector
|
||||
# Now compute the daily adjusted market value of our shares
|
||||
# This is a vector * scalar which yields another vector. Put
|
||||
# another way, we are multiplying each row in the adjusted closing
|
||||
# price by the # shares that we hold.
|
||||
|
||||
MSFT.daily.book.value = Ad(MSFT) * MSFT.starting.shares
|
||||
QQQ.daily.book.value = Ad(QQQ) * QQQ.starting.shares
|
||||
MSFT.daily.market.value = Ad(MSFT) * MSFT.starting.shares
|
||||
QQQ.daily.market.value = Ad(QQQ) * QQQ.starting.shares
|
||||
|
||||
# Now let's plot the value of our investments in QQQ vs. MSFT
|
||||
# on a single chart to see which one wins.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(QQQ.daily.book.value, name = "QQQ") %>%
|
||||
hc_add_series(MSFT.daily.book.value, name = "MSFT") %>%
|
||||
hc_add_series(QQQ.daily.market.value, name = "QQQ") %>%
|
||||
hc_add_series(MSFT.daily.market.value, name = "MSFT") %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# Now let's compute the book value based on the amount invested in a stock
|
||||
# Now let's compute the market value based on the amount invested
|
||||
# in an arbitrary stock. We'll define a new function,
|
||||
# compute.daily.market.value(), which takes a symbol, the amount
|
||||
# invested in it and returns a list that contains the market
|
||||
# value for each day in the investment range.
|
||||
|
||||
compute.daily.book.value <- function(symbol, dollars) {
|
||||
symbol.data <- getSymbols(symbol, auto.assign = FALSE, from = start.date)
|
||||
# There are some new notable R features in this example:
|
||||
# 1. Functions are defined as values
|
||||
# 2. The return statement returns the result of an expression. If
|
||||
# you don't use return, then the value of the last evaluated
|
||||
# expression will be returned from your function.
|
||||
|
||||
# When you run the function, you can either:
|
||||
# 1. Select all the code in the editor and press CTRL+ENTER to
|
||||
# send it in one chunk to the REPL.
|
||||
# 2. Send one line at a time by repeatedly pressing CTRL+ENTER
|
||||
# to see our multi-line support in the REPL. You'll see this
|
||||
# in the "+" characters running down the left margin to
|
||||
# indicate that you're continuing the current block of code
|
||||
# in the REPL.
|
||||
|
||||
compute.daily.market.value <- function(symbol, dollars) {
|
||||
symbol.data <- getSymbols(
|
||||
symbol,
|
||||
auto.assign = FALSE,
|
||||
from = start.date)
|
||||
shares <- dollars / first(Op(symbol.data))
|
||||
return(Ad(symbol.data) * shares)
|
||||
}
|
||||
|
||||
# Now, let's get the daily returns, expressed as a percentage
|
||||
# gain or loss, based on the adjusted prices
|
||||
# gain or loss, based on the adjusted prices. All we have to
|
||||
# do now is call the compute.daily.market.value() with the
|
||||
# name of the symbol, and the starting investment.
|
||||
|
||||
QQQ.daily.book.value <- compute.daily.book.value("QQQ", starting.investment)
|
||||
View(as.data.frame(QQQ.daily.book.value))
|
||||
QQQ.daily.market.value <- compute.daily.market.value(
|
||||
"QQQ",
|
||||
starting.investment)
|
||||
|
||||
# Let's test it on TSLA
|
||||
View(as.data.frame(QQQ.daily.market.value))
|
||||
|
||||
TSLA.daily.book.value <- compute.daily.book.value("TSLA", starting.investment)
|
||||
View(TSLA.daily.book.value)
|
||||
# Now let's do the same thing for an equal weight portfolio
|
||||
# of FANG (Facebook, Amazon, Netflix, Google). Let's define
|
||||
# a data frame that contains two columns: symbols and percentage
|
||||
# that represent the stock symbol names and their percentage
|
||||
# weight in the portfolio (these ideally add up to 1.0 -
|
||||
# enforcing that is an exercise left to the reader).
|
||||
|
||||
TSLA.cumulative.returns <- cumsum(TSLA.daily.returns * starting.investment)
|
||||
head(TSLA.cumulative.returns)
|
||||
|
||||
# Now let's do the same thing for an equal weight portfolio of FANG
|
||||
# You will notice that there is a peculiar named parameter in
|
||||
# the call to the data.frame() function to define the data frame.
|
||||
# stringsAsFactors is something you'll run into often in R -
|
||||
# by default, R interprets character strings as Factors. These
|
||||
# are roughly analogous to enumeration values in other languages.
|
||||
# You get better performance, but these values are not treated
|
||||
# as strings. This parameter forces the data.frame() function to
|
||||
# treat the list of strings assigned to the symbols column as
|
||||
# strings and not factors.
|
||||
|
||||
portfolio <- data.frame(
|
||||
symbols = c("FB", "AMZN", "NFLX", "GOOG"),
|
||||
percentage = c(0.25, 0.25, 0.25, 0.25),
|
||||
stringsAsFactors = FALSE)
|
||||
|
||||
# We need to adjust our daily book value function to compute against portfolio of stocks
|
||||
# Now we need to change our daily market value function to compute
|
||||
# against portfolio of stocks rather than a single stock.
|
||||
|
||||
compute.portfolio.daily.market.value <- function(portfolio, dollars) {
|
||||
|
||||
# The cbind() function binds two dataframes together by column.
|
||||
# In this case, we're appending a new column, dollars, to
|
||||
# the portfolio dataframe. That column contains the result
|
||||
# of muliplying each row in the percentage column against the
|
||||
# dollars scalar value.
|
||||
|
||||
# Note that there are no explicit loops here. In general, if
|
||||
# you find yourself writing a loop in R, you're probably
|
||||
# doing it wrong :)
|
||||
|
||||
portfolio <- cbind(
|
||||
portfolio,
|
||||
dollars = portfolio$percentage * dollars)
|
||||
|
||||
# We define a variable here in the middle of the body of
|
||||
# the compute.portfolio.daily.market.value() function. Because
|
||||
# R supports lexical scoping, this variable can be manipulated
|
||||
# from within the compute.daily.market.value() function that
|
||||
# is defined immediately below it.
|
||||
|
||||
compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
||||
portfolio <- cbind(portfolio, dollars = portfolio$percentage * dollars)
|
||||
df <- NULL
|
||||
|
||||
compute.daily.book.value <- function(symbol, dollars) {
|
||||
# This is a nested function. It can access variables in
|
||||
# the enclosing scope by using the <<- operator.
|
||||
|
||||
compute.daily.market.value <- function(symbol, dollars) {
|
||||
|
||||
symbol.data <- getSymbols(
|
||||
symbol,
|
||||
|
|
@ -128,46 +210,93 @@ compute.portfolio.daily.book.value <- function(portfolio, dollars) {
|
|||
from = start.date)
|
||||
|
||||
shares <- dollars / as.numeric(first(Op(symbol.data)))
|
||||
book.value <- Ad(symbol.data) * shares
|
||||
market.value <- Ad(symbol.data) * shares
|
||||
|
||||
# Here, we use the <<- operator to either:
|
||||
# 1. Assign a new data frame if one didn't exist already
|
||||
# 2. Append a new column to the existing data frame.
|
||||
|
||||
if (is.null(df)) {
|
||||
df <<- data.frame(book.value)
|
||||
df <<- data.frame(market.value)
|
||||
}
|
||||
else {
|
||||
df <<- cbind(df, data.frame(book.value))
|
||||
df <<- cbind(df, data.frame(market.value))
|
||||
}
|
||||
}
|
||||
|
||||
mapply(compute.daily.book.value, portfolio$symbols, portfolio$dollars)
|
||||
# This is how we iterate over all of the rows in the
|
||||
# portfolio data frame. We pass the symbols column to
|
||||
# the compute.daily.market.value() function by dereferencing
|
||||
# it using the $ operator. Similarly, we pass the dollars
|
||||
# column as well. When this function returns, the df
|
||||
# variable will now hold a new data frame containing the
|
||||
# daily market values of all of the stocks in the portfolio.
|
||||
|
||||
# Compute totals for each day
|
||||
mapply(
|
||||
compute.daily.market.value,
|
||||
portfolio$symbols,
|
||||
portfolio$dollars)
|
||||
|
||||
# Compute totals for each day using the rowSums function.
|
||||
# Here we create a new data frame that contains a column
|
||||
# called Total, which contains the sum of the daily
|
||||
# market values of each of the stock symbols.
|
||||
|
||||
df <- cbind(
|
||||
df,
|
||||
data.frame(Total = rowSums(df)))
|
||||
|
||||
df <- cbind(df, data.frame(Total = rowSums(df)))
|
||||
return(df)
|
||||
}
|
||||
|
||||
portfolio.daily.book.value <- compute.portfolio.daily.book.value(portfolio, starting.investment)
|
||||
# Here's a single function call that computes the daily market
|
||||
# value of our portfolio.
|
||||
|
||||
portfolio.daily.market.value <- compute.portfolio.daily.market.value(
|
||||
portfolio,
|
||||
starting.investment)
|
||||
|
||||
# Let's define another "portfolio", but this time it contains only
|
||||
# a single stock - QQQ which represents the NASDAQ 100 index.
|
||||
|
||||
index <- data.frame(
|
||||
symbols = c("QQQ"),
|
||||
percentage = c(1.0),
|
||||
stringsAsFactors = FALSE)
|
||||
|
||||
index.daily.book.value <- compute.portfolio.daily.book.value(index, starting.investment)
|
||||
index.daily.market.value <- compute.portfolio.daily.market.value(
|
||||
index,
|
||||
starting.investment)
|
||||
|
||||
# Now let's plot the two curves against each other
|
||||
# This loses the date labels - better to remove columns?
|
||||
# Our daily market value results are in a data frame.
|
||||
# We need to convert them to an xts object (time series object)
|
||||
# so that the dates display correctly in the plots. We didn't
|
||||
# need to do this in our previous single-stock calculations
|
||||
# because we modified an existing xts object in-place. In our
|
||||
# new portfolio base calculation, we are computing a new data
|
||||
# frame from scratch, hence the need to convert it to
|
||||
# an xts object using the labels from the data frame that
|
||||
# contain the identity of each trading day.
|
||||
|
||||
convert.totals.dataframe.to.xts <- function(df) {
|
||||
return(xts(df$Total, as.POSIXct(rownames(df))))
|
||||
return(
|
||||
xts(
|
||||
df$Total,
|
||||
as.POSIXct(rownames(df))))
|
||||
}
|
||||
|
||||
# Plot the comparison chart between our portfolio and QQQ.
|
||||
|
||||
highchart(type = "stock") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(index.daily.book.value), name = "QQQ") %>%
|
||||
hc_add_series(convert.totals.dataframe.to.xts(portfolio.daily.book.value), name = "Portfolio") %>%
|
||||
hc_add_series(
|
||||
convert.totals.dataframe.to.xts(index.daily.market.value),
|
||||
name = "QQQ") %>%
|
||||
hc_add_series(
|
||||
convert.totals.dataframe.to.xts(portfolio.daily.market.value),
|
||||
name = "Portfolio") %>%
|
||||
hc_add_theme(hc_theme_538())
|
||||
|
||||
# Now the next step in this is to turn this analysis into an applicatoin
|
||||
# where you can pick the stocks to add to your asset allocation and their
|
||||
# proportions. The constraint is that they must add up to 100% in your allocation.
|
||||
# Hit the compute button to generate the results.
|
||||
# Now the next step in this is to turn this analysis into an
|
||||
# application where you can pick the stocks to add to your
|
||||
# asset allocation and their percentages. See the next
|
||||
# section in the lab manual for how to complete this part of the lab.
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче