Files and Resources for Quant Insights Bootcamp (DAY 1)

exercises.txt

Vectorized Backtesting
----------------------
Retrieve data for Apple (AAPL) from 2005 on until yesterday.
Calculate the SMA42 and SMA252 and store them as new columns.
Generate positiongs based on the relationship between the two SMAs.
Compare the performance of the strategy with the market performance.

qi_algo_notes.txt

QUANT INSIGHTS BOOTCAMP
=======================

WIFI
----
Ft1ch#2016#!

MINICONDA
---------
http://conda.pydata.org/miniconda.html
http://conda.pydata.org/docs/using/envs.html

IPYTHON MAGIC COMMANDS
----------------------
https://ipython.org/ipython-doc/3/interactive/magics.html

QUANT PLATFORM
--------------
http://training.pqp.io
--> follow registration link
[special ports used are 8000 and 5565]

qi_bootcamp_intro.ipynb

StrategyOptimizer.py

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns; sns.set()
from pandas_datareader import data as web

class StrategyOptimizer(object):
    def __init__(self, symbol, start, end, t1, t2):
        self.symbol = symbol
        self.start = start
        self.end = end
        self.t1 = t1
        self.t2 = t2
        self.get_data()

    def get_data(self):
        d = web.DataReader(self.symbol, data_source='yahoo',
                           start=self.start, end=self.end)['Adj Close']
        d = pd.DataFrame(d)
        d.columns = [self.symbol]
        d['returns'] = np.log(d / d.shift(1))
        self.data = d

    def plot_data(self):
        plt.figure()
        self.data[self.symbol].plot(title='%s' % self.symbol)
    
    def plot_returns(self):
        plt.figure()
        self.data['returns'].hist(bins=35)

    def run_strategy(self, plot=False):
        self.data['t1'] = self.data[self.symbol].rolling(self.t1).mean()
        self.data['t2'] = self.data[self.symbol].rolling(self.t2).mean()
        self.data['position'] = np.where(self.data.t1 > self.data.t2, 1, -1)
        self.data['strategy'] = self.data.position.shift(1) * self.data.returns
        if plot is True:
            self.data.dropna()[['returns', 'strategy']].cumsum().apply(np.exp).plot()
        perf = self.data.dropna()[['returns', 'strategy']].cumsum().apply(np.exp).ix[-1]
        return perf

    def generate_heat_meap(self, range_t1, range_t2):
        results = np.zeros((len(range_t1), len(range_t2)))
        for i, t1 in enumerate(range_t1):
            for j, t2 in enumerate(range_t2):
                self.t1 = t1
                self.t2 = t2
                perf = self.run_strategy()
                net_perf = perf.loc['strategy'] - perf.loc['returns']
                results[i, j] = net_perf
        heat_map = pd.DataFrame(results, index=range_t1, columns=range_t2)
        return heat_map

vectorized_backtesting.py

#
# Python for Algorithmic Trading
# Quant Insights Bootcamp
#
# The Python Quants
#
import sys
import numpy as np
from pandas_datareader import data as web

sym = 'AAPL'
start = '2010-1-1'
t1 = 42
t2 = 252
if len(sys.argv) > 1:
    start = sys.argv[1]
    sym = sys.argv[2]
    t1 = int(sys.argv[3])
    t2 = int(sys.argv[4])

data = web.DataReader(sym, data_source='yahoo', start=start)

data['SMAshort'] = data['Adj Close'].rolling(t1).mean()
data['SMAlong'] = data['Adj Close'].rolling(t2).mean()
data['Position'] = np.where(data['SMAshort'] > data['SMAlong'], 1, -1)
data['Returns'] = np.log(data['Adj Close'] / data['Adj Close'].shift(1))
data['Strategy'] = data['Position'].shift(1) * data['Returns']
data.dropna(inplace=True)

data[['Adj Close', 'SMAshort', 'SMAlong']].plot(
    title='%s | SMAs = %d & %d' % (sym, t1, t2))
data[['Returns', 'Strategy']].cumsum().apply(np.exp).plot(title=sym)

zmqClient.py

#
# Python for Algorithmic Trading
# Quant Insights Bootcamp
#
# The Python Quants
#
import zmq
import datetime

context = zmq.Context()
socket = context.socket(zmq.SUB)
socket.connect('tcp://0.0.0.0:6666')
socket.setsockopt_string(zmq.SUBSCRIBE, 'AAPL')

while True:
    data = socket.recv_string()
    print(str(datetime.datetime.now())+ ' | ' + data)

zmqCollector.py

#
# Python for Algorithmic Trading
# Quant Insights Bootcamp
#
# The Python Quants
#
import zmq
import pandas as pd
import datetime

context = zmq.Context()
socket = context.socket(zmq.SUB)
socket.connect('tcp://0.0.0.0:6666')
socket.setsockopt_string(zmq.SUBSCRIBE, 'AAPL')

df = pd.DataFrame()

for i in range(20):
    data = socket.recv_string()
    sym, value = data.split()
    timestamp = datetime.datetime.now()
    df = df.append(pd.DataFrame({sym: value}, index=[timestamp]))
    print(i, data)

print(df)

zmqServer.py

#
# Python for Algorithmic Trading
# Quant Insights Bootcamp
#
# The Python Quants
#
import zmq
import time
import random

context = zmq.Context()
socket = context.socket(zmq.PUB)
socket.bind('tcp://0.0.0.0:6666')

AAPL = 100.

while True:
    AAPL += random.gauss(0, 1)
    out = 'AAPL %f' % AAPL
    print(out)
    socket.send_string(out)
    time.sleep(random.random() * 3)

import zmq
import datetime

context = zmq.Context()
socket = context.socket(zmq.SUB)
socket.connect('tcp://0.0.0.0:6666')
socket.setsockopt_string(zmq.SUBSCRIBE, 'AAPL')

while True:
    data = socket.recv_string()
    print(str(datetime.datetime.now())+ ' | ' + data)

zmqTrends.py

#
# Python for Algorithmic Trading
# Quant Insights Bootcamp
#
# The Python Quants
#
import zmq
import numpy as np
import pandas as pd
import datetime

context = zmq.Context()
socket = context.socket(zmq.SUB)
socket.connect('tcp://0.0.0.0:6666')
socket.setsockopt_string(zmq.SUBSCRIBE, 'AAPL')

df = pd.DataFrame()

for i in range(20):
    data = socket.recv_string()
    sym, value = data.split()
    timestamp = datetime.datetime.now()
    df = df.append(pd.DataFrame({sym: value}, index=[timestamp]))
    if i >= 9:
        df['t1'] = df[sym].rolling(5).mean()
        df['t2'] = df[sym].rolling(10).mean()
        df['position'] = np.where(df.t1 > df.t2, 1, -1)
        # place trading logic here
    print(i, df.ix[-1])

print(df)