Median calculation

carlrom

Hello,

I am new to backtrader and until today had managed to find all the answers in the Community. I can't find help on how to calculate the median (P50) of a line for the last 'period' bars (so basically a movav but for the median not the mean). I am creating an indicator and have managed to implement the calculation (sorting and retrieving the value in the middle) within next() but for efficiency and plotting I would like to have a class for the indicator and the calculation in init.

Many thanks.

run-out

@carlrom
I think I would be inclined to use the statistics package for python and then use their median method.

import statistics

You can create an indictor class using next in the indicator to create your line. It would look like this.

class Median(bt.Indicator):

    lines = ("median", )

    params = (
        ("median_period", 10),
    )

    def __init__(self):
        self.addminperiod(self.p.median_period)

    def next(self):
        close_array = self.data.get(size=self.p.median_period)
        self.l.median[0] = statistics.median(close_array)
        # logging
        print(self.data.datetime.date(), close_array, self.l.median[0])

Use the self.addminperiod to ensure there are no empty list for statistics and that you have the minimum period of data in your median calculation.

Plotting can be controlled according to indicator.plots

Your indicator can be called from your strategy init using:

def __init__(self):
    self.median = Median(median_period=self.p.median_period)

The logging looks like this for median_period=5:

2020-11-06 array('d', [261.36, 265.3, 287.38, 294.68, 293.41]) 287.38
2020-11-09 array('d', [265.3, 287.38, 294.68, 293.41, 278.77]) 287.38
2020-11-10 array('d', [287.38, 294.68, 293.41, 278.77, 272.43]) 287.38
2020-11-11 array('d', [294.68, 293.41, 278.77, 272.43, 276.48]) 278.77
2020-11-12 array('d', [293.41, 278.77, 272.43, 276.48, 275.08]) 276.48

The entire code:

import datetime
import backtrader as bt
import statistics


class Median(bt.Indicator):

    lines = ("median", )

    params = (
        ("median_period", 10),
    )

    def __init__(self):
        self.addminperiod(self.p.median_period)

    def next(self):
        close_array = self.data.get(size=self.p.median_period)
        self.l.median[0] = statistics.median(close_array)
        # Logging
        print(self.data.datetime.date(), close_array, self.l.median[0])


class Strategy(bt.Strategy):

    params = (
        ("median_period", 5),
    )

    def log(self, txt, dt=None):
        """ Logging function fot this strategy"""
        dt = dt or self.data.datetime[0]
        if isinstance(dt, float):
            dt = bt.num2date(dt)
        print("%s, %s" % (dt.date(), txt))

    def print_signal(self):
        self.log(
            f"o {self.datas[0].open[0]:7.2f} "
            f"h {self.datas[0].high[0]:7.2f} "
            f"l {self.datas[0].low[0]:7.2f} "
            f"c {self.datas[0].close[0]:7.2f} "
            f"v {self.datas[0].volume[0]:7.0f} "
            f"median {self.median[0]:5.0f}"
        )


    def __init__(self):
        self.median = Median(median_period=self.p.median_period)

    def next(self):
        self.print_signal()


if __name__ == "__main__":

    cerebro = bt.Cerebro()

    tickers = ['FB']

    for ticker in tickers:
        data = bt.feeds.YahooFinanceData(
            dataname=ticker,
            timeframe=bt.TimeFrame.Days,
            fromdate=datetime.datetime(2020, 11, 1),
            todate=datetime.datetime(2020, 12, 5),
            reverse=False,
        )
        cerebro.adddata(data, name=ticker)

    cerebro.addstrategy(Strategy)

    # Execute
    cerebro.run()

carlrom

@run-out It worked! I had tried running the next() within the indicator class but I wasn't including the subscript [0] in median[0].
THANKS A LOT!