Complete BRISMA Pipeline Visualization
Ultra-comprehensive interactive dashboard with 56 Plotly charts covering the entire Example 1 pipeline from data loading to final covariance matrices.
Pipeline Stages
- 1. Data Loading & Preprocessing (8 charts)
- 2. Returns Analysis (8 charts)
- 3. Iterative Covariance Estimation (10 charts)
- 4. Factor Model & Shrinkage (8 charts)
- 5. GARCH Volatility Forecasting (8 charts)
- 6. Covariance Matrix Comparisons (8 charts)
- 7. Edge Cases & Diagnostics (6 charts)
- 8. Mathematical Appendix (Equations & Citations)
Dashboard Categories
1. Data Loading & Preprocessing
Load and preprocess data from brisma_data.xlsx. Includes currency conversion, excess return calculation, and data quality assessment.
Charts: Raw indices, portfolio indices, FX rates, TWI indices, data coverage
View 8 Charts2. Returns Analysis
Calculate and analyze daily and 22-day rolling returns. Examine distributions, autocorrelations, and cross-sectional patterns.
Charts: Return distributions, rolling returns, autocorrelation, drawdowns
View 8 Charts3. Iterative Covariance Estimation
The core algorithm: iterative GARCH-weighted covariance estimation with eigenvalue decomposition and component extraction.
Charts: Weight evolution, eigenspectra, correlation heatmaps, rotations
View 10 Charts4. Factor Model & Shrinkage
Estimate factor loadings and construct the shrinkage covariance matrix combining systematic and idiosyncratic risk.
Charts: Component selection, beta heatmap, fitted vs empirical, R-squared
View 8 Charts5. GARCH Volatility Forecasting
Fit GARCH(1,1) models to components and residuals for forward-looking volatility forecasts.
Charts: GARCH variances, volatility forecasts, term structure
View 8 Charts6. Covariance Matrix Comparisons
Compare the three output matrices: empirical, shrinkage, and GARCH. Analyze eigenvalues, condition numbers, and correlation differences.
Charts: Volatility scatters, correlation diffs, eigenvalue comparison
View 8 Charts7. Edge Cases & Diagnostics
Diagnostics for edge cases: missing data, near-singular matrices, portfolio subsetting, and lookback sensitivity.
Charts: Missing data, singularity detection, regime analysis
View 6 Charts8. Mathematical Appendix
Full mathematical derivations, definitions, and academic citations for all algorithms used in the pipeline.
Content: LaTeX equations, pseudocode, literature references
View Appendix