freedomfaighter / efficient-frontiers

Module to grab stock information and use it to calculate a resampled efficient frontier of portfolio compositions for the desired assets.

efficient-frontiers

Module containing functions that grab stock data from Alpha Vantage and use historical returns (or provided expected returns) to create a Resampled Efficient Frontier for the asset portfolio. The idea of resampled frontiers comes from Michaud & Michaud 2008 (https://www.newfrontieradvisors.com/media/1138/estimation-error-and-portfolio-optimization-12-05.pdf) and produces smoother, more conservative portfolio transitions from one risk level to another compared to classical mean-variance efficient frontiers from Markowitz 1956. The portfolio optimization and resampling techniques in this module do not consider portfolio rebalancing or any time-to-trade rules; the portfolios are considered static and trading decisions are left to the user. See the Wikipedia page for more information on portfolio optimization (https://en.wikipedia.org/wiki/Modern_portfolio_theory). The figures below show examples of efficient frontiers (left panel) and resampled portfolio composition as a function of risk (right panel).

Note that the frontier resampler and plot functions can be used without querying Alpha Vantage. This module was written in part to get experience pulling stock data from Alpha Vantage but any array of asset returns, standard deviations, and correlations will work for the efficient frontier calculations.

Functions

get_stock_info

frontier_resampler

ref_plots

Package dependencies

numpy

pandas

alpha_vantage.timeseries

cvxopt

matplotlib