Portfolio API

Portfolio analytics operate on multi-symbol DataFrames with timestamp, symbol, and a price column such as close.

fs.portfolio.correlation_matrix(df, column="close", method="pearson")

Returns a symbol-by-symbol return correlation matrix. method supports "pearson" and "spearman".

fs.portfolio.covariance_matrix(df, column="close", annualize=True)

Returns a symbol-by-symbol return covariance matrix. When annualize=True, values are multiplied by 252 trading days.

fs.portfolio.rolling_correlation(df, symbol_a, symbol_b, window=63, column="close")

Returns pairwise rolling return correlation for two symbols.

fs.portfolio.portfolio_returns(df, weights, column="close")

Returns weighted portfolio simple returns by timestamp.

portfolio = fs.portfolio.portfolio_returns(df, {"AAPL": 0.5, "MSFT": 0.5})

fs.portfolio.equal_weight_returns(df, column="close")

Returns equal-weight portfolio simple returns.

fs.portfolio.minimum_variance_weights(df, column="close")

Returns minimum-variance weights as a dict[str, float].