Correlation Measure¶

perfana.core.relative.correlation_measure(portfolio, benchmark, duration='monthly', *, is_returns=False, date_as_index=True)[source]¶

Computes the correlation measure through time. The data is assumed to be daily. If the benchmark is a single series, a single TimeSeriesData will be returned. Otherwise, a dictionary of TimeSeries will be returned where the keys are each individual benchmark

Parameters

portfolio (Union[DataFrame, Iterable[Union[int, float]], ndarray, Series]) – The portfolio values vector or matrix
benchmark (Union[DataFrame, Iterable[Union[int, float]], ndarray, Series]) – The benchmark values vector or matrix
duration (Union[str, int]) – Duration to calculate the relative price index with. Either a string or positive integer value can be specified. Supported string values are ‘day’, ‘week’, ‘month’, ‘quarter’, ‘semi-annual’ and ‘year’
is_returns – Set this to true if the portfolio and benchmark values are in “returns” instead of raw values (i.e. prices or raw index value)
date_as_index – If true, returns the date as the dataframe’s index. Otherwise, the date is placed as a column in the dataframe

Returns

A DataFrame of the correlation measure between the assets in the portfolio against the benchmark If multiple series are included in the benchmark, returns a dictionary where the keys are the benchmarks’ name and the values are the correlation measure of the portfolio against that particular benchmark

Return type

TimeSeriesData or dict of TimeSeriesData

Examples

>>> from perfana.datasets import load_etf
>>> from perfana.core import correlation_measure
>>> etf = load_etf().dropna()
>>> returns = etf.iloc[:, 1:]
>>> benchmark = etf.iloc[:, 0]
>>> correlation_measure(returns, benchmark, 'monthly').head()
                 BND       VTI       VWO
Date
2007-05-10 -0.384576  0.890783  0.846000
2007-05-11 -0.525299  0.911693  0.857288
2007-05-14 -0.482180  0.912002  0.855114
2007-05-15 -0.439073  0.913992  0.842561
2007-05-16 -0.487110  0.899859  0.837781