This repo is for demonstration purposes only. It’s a stripped-down snapshot from a private repository, so some proprietary or more advanced techniques have been omitted or replaced with simpler stubs. Please note that this is not meant to be a fully advanced or production-ready codebase.

├── plotting                         # Plotting helpers
│   ├── factors_modelval.py
│   ├── factors.py
│   ├── portfolio_opt.py
│   └── regime_clustering.py
│
├── factor_model                     # Proprietary (intro samples only)
│   ├── fundamental
│   │   ├── exposure_builder.py       
│   │   ├── factor_definitions.py     
│   │   └── fundamental_fitter.py     
│   ├── statistical_fitter.py         
│   └── factor_model.py               
│
├── portfolio_opt                    # Proprietary (intro samples only)
│   ├── constraints
│   │   ├── notional
│   │   │   ├── basket_notional_constraint.py
│   │   │   └── turnover_constraint.py
│   │   ├── base.py
│   │   ├── setup_constraint.py
│   │   └── structured_constraint.py
│   ├── objectives
│   │   ├── base.py
│   │   ├── factor_risk_objective.py
│   │   └── linear_objective.py
│   ├── settings.py       
│   └── cvxpy_wrapper.py              
│
├── regimes                          # Proprietary (backend code not included, more on https://github.com/andreab0/demo-tsrd)
│   ├── helpers
│   │   └── windowing.py             
│   └── utils.py                     
│
├── signals                          # A sample momentum signal
│   └── momentum.py                  
│
├── Sample 1 - Factor Models.ipynb
├── Sample 2 - Regime Clustering.ipynb
│
└── README.md

1. Factor Model
   Demonstrates how to build both fundamental and statistical factor risk models.

   • Fundamental approach: uses categorical or numeric exposures, then fits factor returns via simple OLS or Huber regression.
   • Statistical approach: a basic iterative SVD-based model to estimate factors.

2. Portfolio Optimization
   Includes a lightweight CVXPY wrapper and sample constraints/objectives (e.g., turnover constraints, factor risk minimization).

   • This is only a demo—actual advanced constraints or solvers are omitted or stubbed.

3. Regime Detection
   Illustrates clustering-based regime identification using placeholders for Earth Mover’s Distance (EMD) and K-means.
   An additional demo-repo with a collection of sophisticated regime detection models is available here . Please message me for access.

   • For background, see Earth Mover's Distance and K-means clustering.
     More literature on the subject can be found here.

4. Signals
   A toy momentum signal generator (momentum.py).

• Not Advanced: This code is deliberately simplified and is not intended as a robust, production-grade system.
• Partial Extraction: Certain proprietary algorithms and advanced logic are not included (using serialized files instead).
• For Demo Only: Provided as a reference for exploring basic factor modeling, optimization, and clustering concepts.

• CVXPY Documentation – For modeling and solving convex optimization problems.
• Convex Optimization by Boyd & Vandenberghe – Foundational text on convex optimization techniques.
• Fama-French Factor Models – Introduction to popular fundamental factor models.
• Principal Component Analysis – Basis for the statistical factor model fitter using SVD.
• Robust Regression (HuberRegressor) – Technique for robust regression used in factor model fitting.
• Earth Mover’s Distance – Measure used in regime clustering to compare distributions.
• K-means Clustering – Standard algorithm for partitioning data into clusters.