Features

• Performance indicators calculation

• Supports single time series data and multiple time series data with different time spans

• Time series visualization

Installation

You can install “performance-analysis” via ‘pip’_ from ‘PyPI’_:

$ pip install performance-analysis

Usage

• Performance part

from performance_analysis.Performance import Performance
# Input return data
raw_return_data = pd.read_csv("./raw_return_data.csv")
# Just some examples. For more functions, you can explore the package
ann_rtn = Performance.annualized_return(raw_return_data, period = Constant.DAILY, logreturn = False)
var = Performance.value_at_risk(raw_return_data, significance_level = 0.05)
sharpe = Performance.sharpe_ratio(raw_return_data, risk_free = 0., logreturn = False)
calmar = Performance.calmar_ratio(raw_return_data, period = Constant.DAILY, logreturn = False)

• Computes personal specified indicators

'''
indicators = {
        0 : annualized_return,
        1 : annualized_sd,
        2 : max_drawdown,
        3 : sharpe_ratio,
        4 : calmar_ratio,
        5 : burke_ratio,
        6 : downside_risk,
        7 : sortino_ratio,
        8 : tracking_error,
        9 : information_ratio,
        10 : capm_beta,
        11 : capm_alpha,
        12 : treynor_ratio,
        13 : skewness,
        14 : kurtosis,
        15 : value_at_risk,
        16 : conditional_value_at_risk,
        17 : omega_ratio,
        18 : tail_dependence,
        19 : TDC,
    }
'''

args = (0,1,2,3,4)
kwargs = {
    "annualized_return" : {"returns" : single_return_data},
    "annualized_sd" : {"returns" : single_return_data},
    "max_drawdown" : {"returns" : single_return_data},
    "sharpe_ratio" : {"returns" : single_return_data},
    "calmar_ratio" : {"returns" : single_return_data}
}
performance = Performance.performance_dashboard(*args, **kwargs)

• Plotting part

from performance_analysis.Plotting import Plotting
# read data, set index and convert to datatime
single_return_data = pd.read_csv("./single_return_data.csv")
single_return_data.set_index(['Date'],inplace=True)
single_return_data.index = pd.to_datetime(single_return_data.index, format='%Y%m%d', errors='coerce')

Plotting.plot_cum_return_and_drawdown(single_return_data)
Plotting.plot_monthly_return_heatmap(single_return_data, show_text = True)
Plotting.plot_seasonal_effect(single_return_data)

License

Distributed under the terms of the ‘MIT’_ license, “performance-analysis” is free and open source software