Screen stocks based on Bollinger bands

Question

I am trying to screen several stocks in a list based on Bollinger bands.

Let's say this is the list: c("AMZN","GOOG","TSLA","AAPL")

This is how I am trying to start the code:

stock.list = c("AMZN","GOOG","TSLA","AAPL")
res = list()
for(ss in stock.list) {
stock.data = na.omit(getSymbols(ss, from="1900-01-01", auto.assign=F))
}

I am not sure how to proceed further, but I want to filter the stocks in the list to select only those below the 2nd standard deviation for 50-day average: addBBands(n=50, sd=2)

How would I do that?

Answer 1

Extract the closing prices to a single xts data frame and use rollapply from the zoo package to calculate the bands. Then its smooth sailing.

library(quantmod)

stock.list = c("AMZN","GOOG","TSLA","AAPL")

res = list()

# Modifying your loop to store closing prices in a separate data frame
df = xts()
for(ss in stock.list) {
  res[[ss]] <-na.omit(getSymbols(ss, from="1900-01-01", auto.assign=F))
  names(res[[ss]]) <- c("Open", "High", "Low", "Close", "Volume", "Adjusted")
  df <- cbind(df, res[[ss]][, "Close"])
}

names(df) <- stock.list


# Calculate standard deviation and moving average
sigma <- rollapply(df,
                   width = 50,
                   FUN = sd,
                   by.column = TRUE,
                   align = "right",
                   fill = NA)

mu <- rollapply(df,
                width = 50,
                FUN = mean,
                by.column = TRUE,
                align = "right",
                fill = NA)


# Calculate bollinger bands
upperBand <- mu + 1.96 * sigma
lowerBand <- mu - 1.96 * sigma

# Detect signals
breachedUpper <- df > upperBand
breachedLower <- df < lowerBand

# AAPL has breached its upper band
tail(breachedUpper)