Event profiler

Inspired in QSTK (http://wiki.quantsoftware.org/index.php?title=QSTK_Tutorial_9), the eventprofiler module is a tool to analyze, statistically, how events affect future equity prices. The event profiler scans over historical data for a specified event and then calculates the impact of that event on the equity prices in the past and the future over a certain lookback period.

The goal of this tool is to help you quickly validate an idea, before moving forward with the backtesting process.

class pyalgotrade.eventprofiler.Results(eventsDict, lookBack, lookForward)

Bases: object

Results from the profiler.

getEventCount()

Returns the number of events occurred. Events that are on the boundary are skipped.

class pyalgotrade.eventprofiler.Predicate

Bases: object

Base class for event identification. You should subclass this to implement the event identification logic.

eventOccurred(instrument, bards)

Override (mandatory) to determine if an event took place in the last bar (bards[-1]).

Parameters:

• instrument (string.) – Instrument identifier.
• bards (pyalgotrade.dataseries.bards.BarDataSeries.) – The BarDataSeries for the given instrument.

Return type:

boolean.

class pyalgotrade.eventprofiler.Profiler(predicate, lookBack, lookForward)

Bases: object

This class is responsible for scanning over historical data and analyzing returns before and after the events.

Parameters:

• predicate (Predicate.) – A Predicate subclass responsible for identifying events.
• lookBack (int.) – The number of bars before the event to analyze. Must be > 0.
• lookForward (int.) – The number of bars after the event to analyze. Must be > 0.

getResults()

Returns the results of the analysis.

Return type:Results.

run(feed, useAdjustedCloseForReturns=True)

Runs the analysis using the bars supplied by the feed.

Parameters:

• barFeed (pyalgotrade.barfeed.BarFeed.) – The bar feed to use to run the analysis.
• useAdjustedCloseForReturns (boolean.) – True if adjusted close values should be used to calculate returns.

pyalgotrade.eventprofiler.plot(profilerResults)

Plots the result of the analysis.

Parameters:profilerResults (Results.) – The result of the analysis

Example

The following example is inspired on the ‘Buy-on-Gap Model’ from Ernie Chan’s book: ‘Algorithmic Trading: Winning Strategies and Their Rationale’:

• The idea is to select a stock near the market open whose returns from their previous day’s lows to today’s open are lower that one standard deviation. The standard deviation is computed using the daily close-to-close returns of the last 90 days. These are the stocks that “gapped down”.
• This is narrowed down by requiring the open price to be higher than the 20-day moving average of the closing price.

from pyalgotrade import eventprofiler
from pyalgotrade.technical import stats
from pyalgotrade.technical import roc
from pyalgotrade.technical import ma
from pyalgotrade.tools import yahoofinance

# Event inspired on an example from Ernie Chan's book:
# 'Algorithmic Trading: Winning Strategies and Their Rationale'


class BuyOnGap(eventprofiler.Predicate):
    def __init__(self, feed):
        stdDevPeriod = 90
        smaPeriod = 20
        self.__returns = {}
        self.__stdDev = {}
        self.__ma = {}
        for instrument in feed.getRegisteredInstruments():
            priceDS = feed[instrument].getAdjCloseDataSeries()
            # Returns over the adjusted close values.
            self.__returns[instrument] = roc.RateOfChange(priceDS, 1)
            # StdDev over those returns.
            self.__stdDev[instrument] = stats.StdDev(self.__returns[instrument], stdDevPeriod)
            # MA over the adjusted close values.
            self.__ma[instrument] = ma.SMA(priceDS, smaPeriod)

    def __gappedDown(self, instrument, bards):
        ret = False
        if self.__stdDev[instrument][-1] is not None:
            prevBar = bards[-2]
            currBar = bards[-1]
            low2OpenRet = (currBar.getOpen(True) - prevBar.getLow(True)) / float(prevBar.getLow(True))
            if low2OpenRet < (self.__returns[instrument][-1] - self.__stdDev[instrument][-1]):
                ret = True
        return ret

    def __aboveSMA(self, instrument, bards):
        ret = False
        if self.__ma[instrument][-1] is not None and bards[-1].getOpen(True) > self.__ma[instrument][-1]:
            ret = True
        return ret

    def eventOccurred(self, instrument, bards):
        ret = False
        if self.__gappedDown(instrument, bards) and self.__aboveSMA(instrument, bards):
            ret = True
        return ret


def main(plot):
    instruments = ["AA", "AES", "AIG"]
    feed = yahoofinance.build_feed(instruments, 2008, 2009, ".")

    predicate = BuyOnGap(feed)
    eventProfiler = eventprofiler.Profiler(predicate, 5, 5)
    eventProfiler.run(feed, True)

    results = eventProfiler.getResults()
    print "%d events found" % (results.getEventCount())
    if plot:
        eventprofiler.plot(results)

if __name__ == "__main__":
    main(True)

The code is doing 4 things:

1. Declaring a Predicate that implements the ‘Buy-on-Gap Model’ event identification.
2. Loading bars for some stocks.
3. Running the analysis.
4. Plotting the results.

This is what the output should look like:

2013-09-20 23:30:45,340 yahoofinance [INFO] Creating data directory
2013-09-20 23:30:45,341 yahoofinance [INFO] Downloading AA 2008 to data/AA-2008-yahoofinance.csv
2013-09-20 23:30:46,092 yahoofinance [INFO] Downloading AES 2008 to data/AES-2008-yahoofinance.csv
2013-09-20 23:30:46,683 yahoofinance [INFO] Downloading AIG 2008 to data/AIG-2008-yahoofinance.csv
2013-09-20 23:30:47,260 yahoofinance [INFO] Downloading AA 2009 to data/AA-2009-yahoofinance.csv
2013-09-20 23:30:48,019 yahoofinance [INFO] Downloading AES 2009 to data/AES-2009-yahoofinance.csv
2013-09-20 23:30:48,761 yahoofinance [INFO] Downloading AIG 2009 to data/AIG-2009-yahoofinance.csv
14 events found

Note that Cummulative returns are normalized to the time of the event.