Real Data Dashboard

BRISMA Analysis with Actual Portfolio Data from brisma_data.xlsx

0 Portfolio Assets
0 Risk Factors
0 Observations
0 Years History
View Summary Dashboard Compare with Simulated

Data Overview

Data Source

Real portfolio data from brisma_data.xlsx including daily returns, portfolio weights, risk factors, and yield curve data.

Coverage Period

10+ years of historical data (2558 daily observations) enabling robust covariance estimation and regime analysis.

Asset Universe

31 diversified assets across equities, fixed income, commodities, and currencies with 61 associated risk factors.

Methodology

Full BRISMA pipeline: iterative covariance, GARCH forecasting, Lambda M1/M2 calibration, and hybrid implied returns.

All 19 Real Data Charts

Time Series

01. Portfolio Time Series

Daily returns for all 31 portfolio assets
Time Series

02. Risk Model Time Series

Factor returns driving systematic risk
Time Series

03. Cumulative Returns

Wealth evolution for each asset
Covariance

04. Correlation Heatmap

Asset correlation matrix visualization
Covariance

05. Rolling Volatility

Time-varying volatility estimates
Covariance

14. Covariance Comparison

Empirical vs shrinkage vs GARCH
Portfolio

06. Portfolio Weights

Asset allocation over time
Portfolio

07. Risk Contribution

Marginal risk contribution per asset
Lambda

08. Yield Curve Data

Term structure used for Lambda M1
Lambda

09. Lambda M1 Time Series

Market-based risk aversion estimate
Lambda

10. Lambda M2 Time Series

Historical factor-based lambda
Lambda

11. R-Squared Time Series

Factor model fit for regime switching
Lambda

12. Lambda Comparison

M1 vs M2 side-by-side analysis
Lambda

16. Time Weights

Exponential decay weights for M2
Factors

13. Factor Loadings

Asset exposure to systematic factors
Factors

18. PCA Components

Principal components of return covariance
Implied Returns

17. Implied Returns

mu* from inverse optimization
Implied Returns

15. Efficient Frontier

Mean-variance optimal portfolios
Dashboard

19. Summary Dashboard

Complete overview of all metrics

Live Dashboard Preview

Real Data Summary Dashboard

Open Full Screen

Methodology Applied

Lambda M1 (Market-Based)

lambda = ln((1+y_10Y)/(1+r_f)) / beta_ref

Calibrated from 10-year yield spread and reference beta

Lambda M2 (Historical)

lambda = sum(w_t * f_t)

Exponentially weighted historical factor returns

Implied Returns

mu* = lambda * Q * w + r_f

Expected returns from inverse optimization

Hybrid Blend

mu = R^2 * mu_M1 + (1-R^2) * mu_M2

Regime-adaptive blending based on factor model fit