[R-SIG-Finance] quantmod Charting

[Jeff Ryan]

Some details on how quantmod/chartSeries works are in order:

Looping (or calling) outside of the .GlobalEnv is tricky.  To make
chartSeries and addTA stuff  interactively they do a lot of internal
checking and manipulation.

Calling addTA (or add***) from a non-global env will cause the plot
step to be skipped. To force plotting you need to wrap the add calls
in plot()

Secondly, opening a device first, then looping over these values will
cause a full screen redraw at each iteration -- usually what you want.
 In a pdf() this causes a new plot to be printed.  So if you call
addTA() 10 times on a chart, you'll have a total of 11 charts/pages.

The best way (at present) to manage this is to simply dump the final
chart (after all the additions) to a pdf using:

?saveChart or ?dev.copy2pdf

There are some further nuances to this of course, but that should get
you at least in the right direction.

HTH
Jeff

On Thu, Jul 23, 2009 at 10:53 AM, Brian G. Peterson<brian at braverock.com> wrote:
> You haven't sent us a data series that we can reproduce your problem with.
>  Maybe grab one from an instrument that you don't trade, or simulate a
> series that has some of the same properties that you use to trade.
>
> I typically have no problem using chartSeries on a day of tick data
> (millions of points) or 15-sec OHLC bars (hundreds of thousands of points)
> beyond the time it takes to actually display/render the chart.  This
> includes calling addTA to add extra things to the charts in question.
>
> One thing that comes immediately to mind is that I would probably separate
> the process of creating the indicators from the process of charting.  Run
> through your data series and generate your indicators, then again to
> generate your signals, and *then* chart it from the aligned time series.
>
> There have been multiple previous posts to this list regarding calls to
> chartSeries and addTA from within custom functions, perhaps a review of the
> list archives is in order as well.
>
> Regards,
>
>  - Brian
>
> James Klingsporn wrote:
>>
>> Hello,
>>
>> I am using quantmod to create charts that have buy/sell date times
>> with a window 40 minutes before and after the buy/sell date time.  I
>> want to create ps charts and ultimately pdf charts of the trades.  I'm
>> not concerned with the exit right now.  The number of buy/sell date
>> times can range between 150 – 3000+, so I'd clearly like to automate
>> the process.  The code I’ve been using is attached below.  I've tried
>> putting the below into a function so I call the function with the
>> trade number, but this does not produce the full chart (and a similar
>> problem with 'if' and thus the multipe 'if's below for the same
>> condition).  If I run the below as Source R code it also doesn't
>> produce the full chart.  Also, because the code is so long for each
>> chart, I can't just run the code for all the trades at 1 time; I have
>> to cut and paste the script for about 75 trades/time.  The code has
>> run time of about 5-6 trade charts/min which clearly takes a while for
>> the 3000+ trades and then considering I have also to cut and paste the
>> code.
>>
>> Is there any way to speed up or otherwise improve the script/charting?
>>
>> Thanks,
>> Jim
>>
>>
>> R code for charts -
>>
>>  require(quantmod)
>>
>>        file = 'C:/Consulting/chart.txt'
>>        chart_data <- read.zoo(file, header = FALSE, col.names = c("Date",
>> "Open", "High", "Low", "Close", "MACD1min", "MACD5min", "MACD10min",
>> "Hist1min", "Hist5min", "Hist10min"), sep = ",", dec = ".", tz = "",
>> format = "%Y-%m-%d %H:%M:%S")
>>    ES <- cbind(chart_data$Open, chart_data$High, chart_data$Low,
>> chart_data$Close)
>>        ES <- as.quantmod.OHLC(ES, col.names = c("Open", "High", "Low",
>> "Close"))
>>
>>    time_before = 40
>>    time_after = 40
>>    time_before <- time_before * 60
>>    time_after <- time_after * 60
>>
>>    enter_times_path = 'C:/Consulting/ES_trades_with_slope_rule.txt'
>>    enter_times <- read.zoo(enter_times_path, header = FALSE,
>> col.names = c("Date", "Position"), sep = ",", dec = ".", tz = "",
>> format = "%Y-%m-%d %H:%M:%S")
>>
>>        ctheme <- chartTheme(theme = 'white', area = 'white', up.col =
>> '#0C481A', dn.col = '#B12525')
>>
>> i = 1
>> # Trade number repeats for number of trades
>>
>> display_range <- window(ES, start = time(enter_times[i]) - time_before
>> , end = time(enter_times[i]) + time_after)
>> m <- window(chart_data, start = time(enter_times[i]) - time_before ,
>> end = time(enter_times[i]) + time_after)
>> MACD <- cbind(m$MACD1min, m$MACD5min, m$MACD10min)
>>
>> chartSeries(display_range, theme = ctheme)
>>
>> Hist_range_5 = c(min(range(m$Hist5min)[1], - 0.2),
>> max(range(m$Hist5min)[2], 0.2))
>> Hist_range_10 = c(min(range(m$Hist10min)[1], - 0.2),
>> max(range(m$Hist10min)[2], 0.2))
>>
>> h_steep_slope_1min <- m[abs(diff(m$Hist1min)) >= 0.1]
>> h_slope_1min <- m[abs(diff(m$Hist1min)) >= 0.05 & abs(diff(m$Hist1min)) <
>> 0.1]
>> h_no_slope_1min <- m[abs(diff(m$Hist1min)) < 0.05]
>> h_slope_macd_above_1min <- h_slope_1min[abs(h_slope_1min$Hist1min) > 0.2,]
>> h_slope_macd_below_1min <- h_slope_1min[abs(h_slope_1min$Hist1min) <=
>> 0.2,]
>> h_no_slope_above_1min <- h_no_slope_1min[abs(h_no_slope_1min$Hist1min) >
>> 0.2,]
>> h_no_slope_below_1min <- h_no_slope_1min[abs(h_no_slope_1min$Hist1min) <=
>> 0.2,]
>> h_steep_slope_above_1min <-
>> h_steep_slope_1min[abs(h_steep_slope_1min$Hist1min) > 0.2,]
>> h_steep_slope_below_1min <-
>> h_steep_slope_1min[abs(h_steep_slope_1min$Hist1min) <= 0.2,]
>>
>> a = h_slope_macd_below_1min$Hist1min
>> b = h_no_slope_below_1min$Hist1min
>> f =     h_steep_slope_below_1min$Hist1min
>>
>> c = h_steep_slope_above_1min$Hist1min
>> d = h_no_slope_above_1min$Hist1min
>> e = h_slope_macd_above_1min$Hist1min
>>
>> zero_times = time(m)
>> zero = zoo(0, order.by = zero_times)
>>
>> addTA(a, on = NA, col = '#C03E15' , type = 'h', lwd = 3, yrange =
>> range(m$Hist1min))
>> addTA(b, on = 2, col = '#C03E15', type = 'h', lwd = 1)
>> addTA(f, on = 2, col = '#C03E15', type = 'h', lwd = 6)
>> addTA(c, on = 2, col = '#060440', type = 'h', lwd = 3)
>> addTA(d, on = 2, col = '#060440', type = 'h', lwd = 1)
>> addTA(e, on = 2, col = '#060440', type = 'h', lwd = 6)
>> addTA(zero, on = 2, col = 'black', type = 'l', lwd = 1)
>>
>> h_steep_slope_5min <- m[abs(diff(m$Hist5min)) >= 0.1]
>> h_slope_5min <- m[abs(diff(m$Hist5min)) >= 0.05 & abs(diff(m$Hist5min)) <
>> 0.1]
>> h_no_slope_5min <- m[abs(diff(m$Hist5min)) < 0.05]
>>
>> h_slope_macd_above_5min <- h_slope_5min[abs(h_slope_5min$Hist5min) > 0.7,]
>> h_slope_macd_below_5min <- h_slope_5min[abs(h_slope_5min$Hist5min) <=
>> 0.7,]
>>
>> h_no_slope_above_5min <- h_no_slope_5min[abs(h_no_slope_5min$Hist5min) >
>> 0.7,]
>> h_no_slope_below_5min <- h_no_slope_5min[abs(h_no_slope_5min$Hist5min) <=
>> 0.7,]
>>
>> h_steep_slope_above_5min <-
>> h_steep_slope_5min[abs(h_steep_slope_5min$Hist5min) > 0.7,]
>> h_steep_slope_below_5min <-
>> h_steep_slope_5min[abs(h_steep_slope_5min$Hist5min) <= 0.7,]
>>
>> a5 = h_slope_macd_below_5min$Hist5min
>> b5 = h_no_slope_below_5min$Hist5min
>> f5 = h_steep_slope_below_5min$Hist5min
>>
>> c5 = h_steep_slope_above_5min$Hist5min
>> d5 = h_no_slope_above_5min$Hist5min
>> e5 = h_slope_macd_above_5min$Hist5min
>>
>> addTA(a5, on = NA, col = '#C03E15' , type = 'h', lwd = 3, yrange =
>> range(Hist_range_5))
>> addTA(b5, on = 3, col = '#C03E15', type = 'h', lwd = 1)
>> addTA(f5, on = 3, col = '#C03E15', type = 'h', lwd = 6)
>> addTA(c5, on = 3, col = '#060440', type = 'h', lwd = 3)
>> addTA(d5, on = 3, col = '#060440', type = 'h', lwd = 1)
>> addTA(e5, on = 3, col = '#060440', type = 'h', lwd = 6)
>> addTA(zero, on = 3, col = 'black', type = 'l', lwd = 1)
>>
>> h_steep_slope_10min <- m[abs(diff(m$Hist10min)) >= 0.1]
>> h_slope_10min <- m[abs(diff(m$Hist10min)) >= 0.05 &
>> abs(diff(m$Hist10min)) < 0.1]
>> h_no_slope_10min <- m[abs(diff(m$Hist10min)) < 0.05]
>>
>> h_slope_macd_above_10min <- h_slope_10min[abs(h_slope_10min$Hist10min) >
>> 0.7,]
>> h_slope_macd_below_10min <- h_slope_10min[abs(h_slope_10min$Hist10min) <=
>> 0.7,]
>>
>> h_no_slope_above_10min <-
>> h_no_slope_10min[abs(h_no_slope_10min$Hist10min) > 0.7,]
>> h_no_slope_below_10min <-
>> h_no_slope_10min[abs(h_no_slope_10min$Hist10min) <= 0.7,]
>>
>> h_steep_slope_above_10min <-
>> h_steep_slope_10min[abs(h_steep_slope_10min$Hist10min) > 0.7,]
>> h_steep_slope_below_10min <-
>> h_steep_slope_10min[abs(h_steep_slope_10min$Hist10min) <= 0.7,]
>>
>> a10 = h_slope_macd_below_10min$Hist10min
>> b10 = h_no_slope_below_10min$Hist10min
>> f10 = h_steep_slope_below_10min$Hist10min
>>
>> c10 = h_steep_slope_above_10min$Hist10min
>> d10 = h_no_slope_above_10min$Hist10min
>> e10 = h_slope_macd_above_10min$Hist10min
>>
>> addTA(d10, on = NA, col = '#060440', type = 'h', lwd = 1)
>> addTA(a10, on = 4, col = '#C03E15' , type = 'h', lwd = 3, yrange =
>> range(Hist_range_10))
>> addTA(b10, on = 4, col = '#C03E15', type = 'h', lwd = 1)
>> addTA(f10, on = 4, col = '#C03E15', type = 'h', lwd = 6)
>> addTA(c10, on = 4, col = '#060440', type = 'h', lwd = 3)
>>
>> addTA(e10, on = 4, col = '#060440', type = 'h', lwd = 6)
>> addTA(zero, on = 4, col = 'black', type = 'l', lwd = 1)
>>
>> addTA(MACD, on = NA, col = c('blue', 'red', 'green'), type = c('l',
>> 'l', 'l'), lwd = c(4, 2, 1))
>> addTA(zero, on = 5, col = 'black', type = 'l', lwd = 1)
>>
>> if (as.character(enter_times[i]) == "Long")     Buy <- ES[time(ES) ==
>> time(enter_times[i]) + 60]$ES.Open
>> if (as.character(enter_times[i]) == "Long")     addTA(Buy, on = 1, col =
>> 'red', type = 'p', pch = 6, lwd = 2)
>> if (as.character(enter_times[i]) == "Long")     addTA(Buy > 0, on =
>> -(1:5), col = '#FFCAFD', border = NA)
>>
>> if (as.character(enter_times[i]) == "Short")    Short <- ES[time(ES) ==
>> time(enter_times[i]) + 60]$ES.Open
>> if (as.character(enter_times[i]) == "Short")    addTA(Short, on = 1,
>> col = 'blue', type = 'p', pch = 2, lwd = 2)
>> if (as.character(enter_times[i]) == "Short")    addTA(Short > 0, on =
>> -(1:5), col = '#FF7878', border = NA)
>>
>> # Repeats with i = 2...
>>
>> _______________________________________________
>> R-SIG-Finance at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
>> -- Subscriber-posting only.
>> -- If you want to post, subscribe first.
>>
>
>
> --
> Brian G. Peterson
> http://braverock.com/brian/
> Ph: 773-459-4973
> IM: bgpbraverock
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
>



-- 
Jeffrey Ryan
jeffrey.ryan at insightalgo.com

ia: insight algorithmics
www.insightalgo.com