IMPORTING LIBRARIES¶

In [82]:

import pandas
import numpy as np
import matplotlib.pyplot as plt
import datetime
import pandas_datareader
from pandas_datareader import data as pdr
import yfinance
import pypfopt as ppo
from pypfopt import risk_models
from pypfopt import expected_returns
from pypfopt.efficient_frontier import EfficientFrontier
from pypfopt.discrete_allocation import DiscreteAllocation,get_latest_prices

IMPORTING HISTORICAL DATA FROM YAHOO FINANCE (DATA COLLECTION , PORTFOLIO CONSTRUCTION)¶

In [83]:

end=datetime.datetime.now()
start=datetime.datetime(2010,1,1)
portfolio=['AMZN','GOOG','MSFT','NFLX','AAPL','NVDA']
data=yfinance.download(portfolio,start,end)
data.head(10)

[*********************100%%**********************]  6 of 6 completed

Out[83]:

	Adj Close						Close				...	Open				Volume
	AAPL	AMZN	GOOG	MSFT	NFLX	NVDA	AAPL	AMZN	GOOG	MSFT	...	GOOG	MSFT	NFLX	NVDA	AAPL	AMZN	GOOG	MSFT	NFLX	NVDA
Date
2010-01-04	6.487533	6.6950	15.610239	23.522566	7.640000	4.240801	7.643214	6.6950	15.610239	30.950001	...	15.615220	30.620001	7.931429	4.6275	493729600	151998000	78541293	38409100	17239600	80020400
2010-01-05	6.498748	6.7345	15.541497	23.530172	7.358571	4.302728	7.656429	6.7345	15.541497	30.959999	...	15.620949	30.850000	7.652857	4.6050	601904800	177038000	120638494	49749600	23753100	72864800
2010-01-06	6.395381	6.6125	15.149715	23.385763	7.617143	4.330250	7.534643	6.6125	15.149715	30.770000	...	15.588072	30.879999	7.361429	4.6875	552160000	143576000	159744526	58182400	23290400	64916800
2010-01-07	6.383556	6.5000	14.797037	23.142559	7.485714	4.245388	7.520714	6.5000	14.797037	30.450001	...	15.178109	30.629999	7.731429	4.6950	477131200	220604000	257533695	50559700	9955400	54779200
2010-01-08	6.425997	6.6760	14.994298	23.302158	7.614286	4.254562	7.570714	6.6760	14.994298	30.660000	...	14.744733	30.280001	7.498571	4.5900	447610800	196610000	189680313	51197400	8180900	47816800
2010-01-11	6.369309	6.5155	14.971633	23.005745	7.604286	4.194932	7.503929	6.5155	14.971633	30.270000	...	15.055070	30.709999	7.660000	4.6625	462229600	175588000	289597429	68754700	6783700	55661200
2010-01-12	6.296857	6.3675	14.706875	22.853754	7.481429	4.052728	7.418571	6.3675	14.706875	30.070000	...	14.885456	30.150000	7.528571	4.5050	594459600	181926000	194859654	65912100	6330100	62743200
2010-01-13	6.385679	6.4555	14.622441	23.066551	7.708571	4.107774	7.523214	6.4555	14.622441	30.350000	...	14.358431	30.260000	7.612857	4.4475	605892000	214464000	260838034	51863500	14422100	50886800
2010-01-14	6.348693	6.3675	14.691184	23.530172	7.284286	4.043553	7.479643	6.3675	14.691184	30.959999	...	14.542989	30.309999	7.518571	4.4225	432894000	195498000	170239717	63228100	17685500	60852400
2010-01-15	6.242596	6.3570	14.445853	23.454165	7.278571	3.924290	7.354643	6.3570	14.445853	30.860001	...	14.778108	31.080000	7.245714	4.3750	594067600	307530000	218194794	79913200	13031200	81819200

10 rows × 36 columns

GETTING CLOSING PRICES FROM DATA , CREATING A COVARIENCE MATIX AND ALLOTING RANDOM WEIGHTS¶

In [84]:

close_data=data["Adj Close"]
returns=close_data.pct_change()
Mean_Returns=returns.mean()
Cov_Matrix=returns.cov()
weights=np.random.random(6) 
weights /= np.sum(weights)
print("Covarience Matrix is : \n\n",Cov_Matrix,"\n")
print("Weights alloted are : ",weights)

Covarience Matrix is : 

           AAPL      AMZN      GOOG      MSFT      NFLX      NVDA
AAPL  0.000318  0.000181  0.000172  0.000174  0.000170  0.000254
AMZN  0.000181  0.000436  0.000217  0.000193  0.000294  0.000274
GOOG  0.000172  0.000217  0.000298  0.000181  0.000200  0.000249
MSFT  0.000174  0.000193  0.000181  0.000269  0.000179  0.000265
NFLX  0.000170  0.000294  0.000200  0.000179  0.001048  0.000308
NVDA  0.000254  0.000274  0.000249  0.000265  0.000308  0.000805 

Weights alloted are :  [0.07132334 0.28657005 0.14540522 0.27390361 0.20401229 0.01878549]

RUNNING SIMULATION AND CREATING VIUSLIZATION (MODEL DESIGN AND SIMULATION)¶

In [85]:

Simulation_number=20000
Time=500
Mean_M=np.full(shape=(Time,len(weights)),fill_value=Mean_Returns)
Mean_M=Mean_M.T
portfolio_sims=np.full(shape=(Time,Simulation_number),fill_value=0.0)
Initial_Portfolio=25000 #Dollars
for m in range(0, Simulation_number):
    Z = np.random.normal(size=(Time,6))
    L = np.linalg.cholesky(Cov_Matrix)
    dailyReturns = Mean_M + np.inner(L, Z)
    portfolio_sims[:,m] = np.cumprod(np.inner(weights, dailyReturns.T)+1)*Initial_Portfolio
plt.plot(portfolio_sims)
plt.ylabel('Portfolio Value($)')
plt.xlabel('Days')
plt.title('Monte Carlo Simulation')
plt.show()   

GETTING OPTIMIZED PORTFOLIO (OPTIMIZATION)¶

In [88]:

latest_prices = get_latest_prices(close_data)
Discrete_Allocation = DiscreteAllocation(weights, latest_prices,total_portfolio_value=10000)
allocation, leftover = Discrete_Allocation.greedy_portfolio()
print("Optimized portfolio for maximum returns : \n", allocation,"\n")
print("Funds remaining are ",leftover,'Dollar .')
print("Amount spent is ",Initial_Portfolio-leftover," Dollars. ")

Optimized portfolio for maximum returns : 
 {'AMZN': 12, 'AAPL': 10, 'MSFT': 5, 'NFLX': 4, 'NVDA': 4, 'GOOG': 13} 

Funds remaining are  96.52011108398438 Dollar .
Amount spent is  24903.479888916016  Dollars.

ALLOTTING CLEANED WEIGHTS AND GETTING PROJECTION OF PORTFOLIO VALUE (ANALYSIS)¶

In [92]:

sample_cov = risk_models.sample_cov(close_data, frequency=42)
Exp_Ret = expected_returns.capm_return(close_data)
Cov_Matrix_2=risk_models.sample_cov(close_data)
ef = EfficientFrontier(Exp_Ret, Cov_Matrix_2)
weights = ef.max_sharpe()
clean_weights = ef.clean_weights()
print("Clean weights are : \n",dict(clean_weights))
print("\nProjections: ")
ef.portfolio_performance(verbose=True)

Clean weights are : 
 {'AAPL': 0.16667, 'AMZN': 0.16667, 'GOOG': 0.16667, 'MSFT': 0.16667, 'NFLX': 0.16667, 'NVDA': 0.16667}

Projections: 
Expected annual return: 30.8%
Annual volatility: 26.2%
Sharpe Ratio: 1.10

Out[92]:

(0.30773190077275975, 0.261855093611706, 1.09882109530176)