#' --- #' title: "Introduction to time series handling and the Quantmod-Package" #' author: "Dana Uhlig" #' -- #getting a lot of Financial data via quantmod from YahooFinance and other sources #install.packages("quantmod") library(quantmod) #loads also xts, zoo, ttr #loading Dow Jones from yahoo getSymbols(Symbols = "^DJI") ls() str(DJI) head(DJI) tail(DJI) names(DJI) DJI.long = getSymbols(Symbols = "^DJI", from="1995-01-01", auto.assign = FALSE) str(DJI.long) head(DJI.long) tail(DJI.long) names(DJI.long) #more symbols getSymbols(c('AAPL','AMZN','VOW.DE','TSLA'),source="yahoo") str(AAPL) head(AAPL) tail(AAPL) #finding symbols via stockSymbols() #all symbols from NYSE (typ. blue chip companies like AT&T, GM) nyse.symbols = stockSymbols("NYSE") dim(nyse.symbols) head(nyse.symbols) tail(nyse.symbols) #saerching using regular expressions (grepl) (gm.nyse = grep('General Motors',nyse.symbols[,2])) nyse.symbols[gm.nyse,] #-> symbol GM #or alternatively which(grepl('General Motors',nyse.symbols[,2])) getSymbols('GM') #all symbols from Nasdaq (typ. tech startups companies like Apple, Amazon, ..) nasdaq.symbols = stockSymbols("NASDAQ") dim(nasdaq.symbols) head(nasdaq.symbols) tail(nasdaq.symbols) (amazon = grep('Amazon',nasdaq.symbols[,2])) nasdaq.symbols[amazon,] all.symbols = stockSymbols(exchange = c("AMEX","NASDAQ","NYSE")) dim(all.symbols) head(all.symbols) tail(all.symbols) str(all.symbols) table(all.symbols$Sector) #plot closin prices from Apple plot(Cl(AAPL)) plot(AAPL$AAPL.Close) #plot(AAPL) chartSeries(VOW.DE) chartSeries(TSLA,subset = 'last 3 years') chartSeries(TSLA,subset = 'last 1 year') chartSeries(TSLA,subset = 'last 250 days') (vw.returns.2016 = dailyReturn(VOW.DE,subset="2016")) (tsla.returns.2016 = dailyReturn(TSLA,subset="2016")) #usual plots (generic) and calculus plot(DJI) x = DJI$DJI.Close #x is xts object -> plot.xts based on chart_Series plot(x) chart_Series(x) #plain plot plot(as.numeric(x),type="l") plot(time(x),as.numeric(x),type="l") #ret = log(x_t / x_{t-1}) = log(x_t) - log(x_{t-1}) x.ret = diff(log(x)) plot(x.ret) hist(x.ret) #same using quantmod function dji.ret = dailyReturn(DJI) dji.ret.2009 = dailyReturn(DJI,subset="2009") hist(dji.ret.2009) dji.ret.2008.2009 = dailyReturn(DJI,subset="2008::2009") plot(dji.ret.2008.2009) hist(dji.ret.2008.2009) dji.ret.y = periodReturn(DJI, period="yearly") dji.ret.y #same as annualReturn or yearlyReturn dji.ret.a = annualReturn(DJI) dji.ret.a - dji.ret.y (test = allReturns(DJI)) plot(dji.ret) plot(dji.ret, main="DJIA Returns", type="n") lines(dji.ret) ################################################################################################### #the package ggplot2 provides nicer graphics #install.packages("ggplot2") library(ggplot2) ggplot(data = dji.ret, aes(daily.returns)) + geom_histogram(aes(y=..density..)) + geom_density(col="green") + stat_function(fun = dnorm, args=list(mean=mean(dji.ret),sd = sd(dji.ret)), colour = "red") #################################################################################################### #financial charts using quantmod chartSeries(DJI) chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2016::") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2016::2017") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="last 12 months") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="last 30 days") chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017",type = 'candlesticks') chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017",type = 'matchsticks') chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017",type = 'bars') chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017",type = 'line') chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red",subset="2017",type = 'auto') #chart overlays and tecnical indicators: #adding MACD (developed by Gerald Appel) popular price oscillator chartSeries(Cl(DJI), theme = "white", up.col = "green", dn.col = "red", TA=" addMACD(fast = 8, slow = 11, signal = 6); addMACD(fast = 2, slow = 4, signal = 19)") ?chartSeries chartSeries(DJI, theme = "white", up.col = "green", dn.col = "red", TA="addRSI(n = 40); addRSI(n = 14)") ## getSymbols("AAPL") chartSeries(AAPL,subset = '2017') plot(addMACD()) plot(addBBands()) plot(addMACD(fast = 8,slow = 11, signal = 6)) ## getSymbols(Symbols = "^GDAXI") head(GDAXI) #introduction: 1988-07-01, Normierung 31. Dezember 1987 auf 1.000 dax.long = getSymbols(Symbols = "^GDAXI",auto.assign = FALSE, from = "1985-01-01") head(dax.long) chartSeries(GDAXI, name = "Deutscher Aktien Index", theme = "white", up.col = "green", dn.col = "red", subset="last 1 year", TA="addSMA()") chartSeries(GDAXI, name = "Deutscher Aktien Index", theme = "white", up.col = "green", dn.col = "red", subset="last 12 months", TA="addSMA()") chartSeries(GDAXI,theme = "white", up.col = "green", dn.col = "red", subset="last 12 months", TA="addSMA()",type = "line" ) chartSeries(Cl(GDAXI), theme = "white", up.col = "green", dn.col = "red", subset="last 12 months", TA="addSMA()",type = "line" ) chartSeries(GDAXI, theme = "white", up.col = "green", dn.col = "red", subset="last 12 months", TA="addSMA()",type="line") chartSeries(GDAXI, theme = "white", up.col = "green", dn.col = "red", subset="last 12 months", TA="addSMA()",type="line") addSMA(on=1,n = 50, col = "blue") chartSeries(GDAXI, name = "Deutscher Aktien Index", theme = "white", up.col = "green", dn.col = "red", subset="2007:2009", TA="addRSI(n = 40); addRSI(n = 14); addMACD(fast = 8, slow = 11, signal = 6); addMACD(fast = 2, slow = 4, signal = 19)") chartSeries(GDAXI, name = "Deutscher Aktien Index", theme = "white", up.col = "green", dn.col = "red", subset="2007", TA="addRSI(n = 40); addRSI(n = 14); addMACD(fast = 8, slow = 11, signal = 6); addMACD(fast = 2, slow = 4, signal = 19)") ?EMA #### #Bitcoin USD (BTC-USD) getSymbols("BTC-USD") ls() `BTC-USD` #problems with '-' sign in name -> `` plot(`BTC-USD`) #plot(BTC-USD) #chartSeries(BTC-USD) chartSeries(`BTC-USD`) #own var bitcoin_usd = getSymbols("BTC-USD",auto.assign = FALSE) head(bitcoin_usd) chartSeries(bitcoin_usd) chartSeries(bitcoin_usd,subset = 'last 4 months') chartSeries(bitcoin_usd,subset = '2018') chartSeries(bitcoin_usd,subset = 'last 12 months') chartSeries(bitcoin_usd, subset='2017-06::2018-05') chartSeries(bitcoin_usd,theme=chartTheme('white'),subset = 'last 12 months') chartSeries(bitcoin_usd,TA=NULL, subset='last 12 months',theme=chartTheme('white')) #no volume addVo() #add Volume addBBands() addMACD() addCCI() addChVol() addRSI() chartSeries(bitcoin_usd,name = "BTC-USD",TA=c(addVo(),addBBands()), subset='last 12 months') #add volume and Bollinger Bands from TTR addMACD() # add MACD indicator to current chart #more Details see http://www.quantmod.com/examples/charting/ ####################################################################################################################### ### other sources (default = yahoo) # 1. Googgle src doesn't work # getSymbols("GOOG", src = "google") # 2. FRED: Federal Reserve Bank of St. Louis - Economic Research #examples Gross Domestic Product (GDP) https://fred.stlouisfed.org/series/GDP/ #gdp.get = getSymbols("GDP",src = "FRED",auto.assign = FALSE) #str(gdp.get) #ls() getSymbols("GDP",src = "FRED") chartSeries(GDP) #U.S. / Euro Foreign Exchange Rate (DEXUSEU) getSymbols("DEXUSEU",src = "FRED") chartSeries(DEXUSEU,theme="white") chartSeries(1/DEXUSEU,theme="white") # 10-Year Treasury Constant Maturity Rate (DGS10) getSymbols("DGS10", src="FRED") chartSeries(DGS10,theme="white") chartSeries(DGS10,theme="white",subset="2007::") #Civilian Unemployment Rate (UNRATE) getSymbols("UNRATE", src="FRED") chartSeries(UNRATE,theme="white") chartSeries(UNRATE,theme="white",subset="2007::") # 3-Month London Interbank Offered Rate (LIBOR), based on U.S. Dollar (USD3MTD156N) getSymbols("USD3MTD156N", src="FRED") chartSeries(USD3MTD156N,theme="white") chartSeries(USD3MTD156N,theme="white",subset="2007::") #3. OANDA getSymbols("USD/EUR",src="oanda",from="2005-01-01") getSymbols("USD/EUR",src="oanda") chartSeries(USDEUR,theme="white") chartSeries(1/USDEUR,theme="white") chartSeries(DEXUSEU,theme="white",subset = '2017-11-29::2018-05-27') #see https://www.quantmod.com/examples/intro/ ##### https://www.quandl.com/ #install.packages("Quandl") library(Quandl) gdp = Quandl(code = "FRED/GDP") str(gdp) # Import GDP data from FRED as xts gdp_xts = Quandl(code = "FRED/GDP", type = "xts") str(gdp_xts) ## # Import GDP data from FRED as zoo gdp_zoo = Quandl(code = "FRED/GDP", type = "zoo") # Look at the structure of gdp_zoo str(gdp_zoo) ################################################################################# #multi assets - quantmod tickers = c('GLD', 'GDX','USO') getSymbols(tickers, src = 'yahoo', from = '2017-01-01') chartSeries(GDX,theme="white",TA=NULL) chartSeries(GLD,theme="white",TA=NULL) chartSeries(USO,theme="white",TA=NULL) chartSeries(rbind(GLD,GDX,USO), TA=NULL) all = rbind(GLD,GDX,USO) chartSeries(all,TA=NULL) normalise_series <- function(xdat) xdat / coredata(xdat)[1] # Define colour of default chart line to chart_Series in mytheme object # which is passed to chart_Series: mytheme = chart_theme() mytheme$col$line.col = "darkgreen" chart_Series(normalise_series(Cl(GLD)) , theme = mytheme) add_TA(normalise_series(Cl(GDX)), on = 1, col = "red", lty = 3) add_TA(normalise_series(Cl(USO)), on = 1, col = "blue", lty =2) add_TA(RSI(Cl(GLD)), on = NA, col = "darkgreen") add_TA(RSI(Cl(GDX)), on = 2, col = "red", lty = 3) # Or add RSIs on different subpanels to improve readability of charts: add_TA(RSI(Cl(USO)), on = 2, col = "blue", lty = 2) ############################################################################################################################ #creating own time series z = ts(matrix(rt(200 * 8, df = 3), 200, 8), start = c(1961, 1), frequency = 12) plot(z, yax.flip = TRUE) plot(z, axes = FALSE, ann = FALSE, frame.plot = TRUE, mar.multi = c(0,0,0,0), oma.multi = c(1,1,5,1)) title("plot(ts(..), axes=FALSE, ann=FALSE, frame.plot=TRUE, mar..., oma...)") str(z) str(DJI) #xts objects zxts = as.xts(matrix(rt(200 * 8, df = 3), 200, 8),frequency = 12,order.by = seq(as.Date("2008-01-01"),length=200, by="weeks")) head(zxts) str(zxts) data = rt(250, df=4) dates = seq(as.Date("2017-01-01"), length=length(data), by = "days") ts1 = xts(x=data, order.by=dates) head(ts1) tail(ts1) str(ts1) chartSeries(ts1) deltat = 1/250 data2 = rnorm(250, mean = 0.03*deltat, sd = 0.2*deltat) dates2 = seq(as.Date("2017-01-01"), length=length(data2), by = "days") ts2 = xts(x=100*exp(data2), order.by=dates2) chartSeries(ts2) ################################# #install.packages("PerformanceAnalytics") library(PerformanceAnalytics) #Analyse und Darstellung der Portfolioperformance fuer Long charts.PerformanceSummary(dji.ret) chartSeries(dji.ret) chartSeries(cumsum(dji.ret)) table.Drawdowns(dji.ret) table.DownsideRisk(dji.ret) sd(dji.ret) quantile(dji.ret,0.95) ############## ############################################################################################## ####### many textbooks with own packages providing book examples and methods #Quelle: https://www.stat.pitt.edu/stoffer/tsa4/ #http://www.stat.pitt.edu/stoffer/tsa4/tsa4.htm # https://www.springer.com/us/book/9783319524511 #-> ASTSA = Applied Statistical Time series Analysis #install.packages("astsa") library(astsa) ?astsa #Johnson & Johnson Quarterly Earnings plot(jj,type="o",main="Johnson & Johnson Quarterly Earnings") plot(globtemp, type="o", ylab="Global Temperature Deviations") plot(speech) # El Niño and Fish Population (Figure 1.5) # monthly values of an environmental series called the Southern Oscillation Index (SOI) # and associated Recruitment (number of new fish) par(mfrow = c(2,1)) # set up the graphics plot(soi, ylab="", xlab="", main="Southern Oscillation Index") plot(rec, ylab="", xlab="", main="Recruitment") par(mfrow = c(1,1)) # reset up the graphics # fig 1.6 par(mfrow=c(2,1)) ts.plot(fmri1[,2:5], col=1:4, ylab="BOLD", main="Cortex") ts.plot(fmri1[,6:9], col=1:4, ylab="BOLD", main="Thalamus & Cerebellum") par(mfrow = c(1,1)) ##############################################################################################