def generate_optimized_portfolio(symbols,
start_date,
end_date,
investment,
rerun=False):
# get new optimized portfolio and get its value for the date
tickers = get_tickers_for_symbols(symbols, start_date, end_date)
weights = []
for symbol in symbols:
weights.append(1 / float(len(symbols)))
portfolio = Portfolio(tickers, weights, start_date, end_date, investment)
optimized_portfolio = optimize_portfolio(portfolio)
# only keep symbols with 2% or higher weight and only run it once to make sure we dont create an inifinite loop
if not rerun:
# create dict of weights and symbols
symb_weight_dict = dict(
zip(optimized_portfolio.symbols, optimized_portfolio.weights))
symbols_to_keep = []
for item in symb_weight_dict:
if symb_weight_dict[item] > 0.02:
symbols_to_keep.append(item)
else:
pass
if len(symbols_to_keep) < len(symbols):
print 'running optimizer one more time'
optimized_portfolio = generate_optimized_portfolio(
symbols_to_keep, start_date, end_date, investment, True)
return optimized_portfolio
def get_portfolio(symbols,
weights,
start_date,
end_date,
investment,
no_spy=True):
prices_df, prices_df_with_spy = get_prices(symbols, start_date, end_date,
no_spy)
tickers = get_tickers_for_symbols(symbols, start_date, end_date, no_spy)
portfolio = Portfolio(tickers, weights, start_date, end_date, investment)
return portfolio
def optimize_portfolio(portfolio):
symbol_alloc, is_reduced = get_top_optimal_symbols(portfolio.ticker_prices)
if is_reduced:
symbols_used = symbol_alloc.index.values
symbol_alloc, is_reduced = get_top_optimal_symbols(
portfolio.ticker_prices)
symbols_used = symbol_alloc.index.values.tolist()
allocations_used = symbol_alloc['Allocations'].values.tolist()
tickers = get_tickers_for_symbols(symbols_used, portfolio.start_date,
portfolio.end_date)
return Portfolio(tickers, allocations_used, portfolio.start_date,
portfolio.end_date, portfolio.investment)
def run():
start_date = '2016-12-23'
end_date = '2017-01-05'
investment = 100000 # $100,000.00 as starting investment
symbols = ['PGR', 'CCI', 'STZ', 'WYNN', 'DPS']
print 'Data for ' + symbols[0], get_data_for_symbol(symbols[0]).head()
weights = [0.2, 0.2, 0.2, 0.2, 0.2]
optimize = False
# analyse_portfolio(symbols, weights, start_date, end_date, investment, optimize)
# print calculated features and label
prices_df, prices_df_with_spy = get_prices([symbols[0]], start_date,
end_date)
ticker_analysed = TickerAnalysed(symbol=symbols[0],
data_df=prices_df[[symbols[0]]])
print 'Stats for: ', ticker_analysed.symbol
# Original data
store_ticker_analysis(ticker_analysed.df,
ticker_analysed.symbol + '_original')
ticker_analysed.print_df()
# after filling missing data
store_ticker_analysis(ticker_analysed.get_features_df(),
ticker_analysed.symbol + '_fill')
print ticker_analysed.get_features_df()
tickers = get_tickers_for_symbols(symbols, start_date, end_date)
portfolio = Portfolio(tickers, weights, start_date, end_date, investment)
print 'Before Optimization'
print '---------------------------'
print 'Comparing portfoltio to S&P'
print '---------------------------'
compare_to_SP(portfolio)
optimized_portfolio = optimize_portfolio(portfolio)
print '---------------------------'
print 'After Optimization: '
print '---------------------------'
optimized_portfolio.describe()
print '---------------------------'
print 'Comparing portfoltio to S&P'
print '---------------------------'
compare_to_SP(optimized_portfolio)
def display_ticker(symbol, start_date, end_date):
tickers = get_tickers_for_symbols([symbol], start_date, end_date)
ticker = tickers[0]
plot_data(ticker.df, symbol + ' Adj Close Price')