Institutional Quantitative Finance Library
A comprehensive 53-module Quant Engine built in Python, replicating the internal architecture of a Tier-1 Hedge Fund.
Core Capabilities:
HFT Microstructure: Order Flow Imbalance (OFI), Queue Estimation, and Binary Data Parsing.
Advanced Mathematics: Stochastic Calculus (Heston), Graph Theory, and Signal Processing (EMD).
Deep Learning: GANs for synthetic data, LSTMs for forecasting, and BERT for NLP.
Engineering: GPU-accelerated Monte Carlo (CUDA), Distributed Cloud Computing (Ray), and Numba JIT.
Risk: Extreme Value Theory (EVT) and Tail-Dependence Copulas.
Designed for Research, Backtesting, and High-Performance Execution.
| File Name | Type | Description | Status |
|---|---|---|---|
01_trend_following_backtest.py |
Backtest | A Long/Short strategy reacting to Moving Average crossovers. Includes Maximum Drawdown calculation to measure risk. | β Simulation |
02_live_paper_trader.py |
Paper Trading | A virtual trading engine that fetches Live NSE Data (via nselib) and simulates order execution with configurable Slippage Models. |
π’ Real-Time Data / Virtual Trade |
03_iceberg_execution_algo.py |
Execution | Algorithmic logic to split large "Parent Orders" into hidden "Child Orders" to minimize market impact (Institutional Execution). | β Simulation |
04_vwap_analyzer.py |
Analysis | Computes VWAP (Volume Weighted Average Price) and generates a time-weighted execution schedule for institutional buying. | β Simulation |
05_pairs_trading_stat_arb.py |
Quant Strategy | A Statistical Arbitrage model using the Engle-Granger test to find Cointegration between banking stocks (HDFC & ICICI) for Market Neutral strategies. | β Backtest |
| File Name | Type | Description | Status |
|---|---|---|---|
06_risk_manager_var.py |
Risk Mgmt | Calculates Value at Risk (VaR) using Historical Simulation. Essential for daily risk reporting and capital adequacy. | β Analytical Tool |
07_portfolio_optimizer.py |
Portfolio Mgmt | Uses Monte Carlo Simulation to generate the Efficient Frontier and identify the portfolio with the Maximum Sharpe Ratio. | β Optimization |
08_option_pricer_black_scholes.py |
Derivatives | Implements the Black-Scholes Model to calculate theoretical Call/Put prices and Greeks (Delta) for hedging logic. | β Calculator |
09_sentiment_analysis_nlp.py |
Alt Data | A basic Natural Language Processing (NLP) script that scores financial news headlines as Bullish or Bearish. | β NLP Demo |
| File Name | Type | Description | Status |
|---|---|---|---|
10_kalman_filter_trend.py |
Signal Processing | Implements a Kalman Filter to estimate the "true" stock price state from noisy data. Reduces lag compared to standard Moving Averages. | β Math Model |
11_ml_xgboost_predictor.py |
Machine Learning | Uses XGBoost (Gradient Boosting) to predict future price direction (Up/Down) based on volatility and momentum features. | β AI Model |
12_volatility_surface_3d.py |
Derivatives | Generates a 3D Volatility Surface to visualize the "Option Smile" and Term Structure, critical for pricing exotic options. | β 3D Visualization |
13_hft_orderbook_sim.py |
Microstructure | A Limit Order Book (L2) matching engine using Heap data structures. Simulates how HFT algos consume liquidity. | β Simulation |
| File Name | Type | Description | Status |
|---|---|---|---|
14_deep_learning_lstm_forecast.py |
Deep Learning | Uses LSTM (Long Short-Term Memory) Neural Networks to predict time-series data, capturing non-linear patterns that standard regressions miss. | β AI / Keras |
15_rl_trading_agent.py |
Reinforcement Learning | A Q-Learning Agent that learns to trade by trial-and-error. It optimizes a reward function (P&L) rather than predicting prices. | β RL / AI |
16_hierarchical_risk_parity.py |
Portfolio Optimization | Uses Graph Theory & Clustering to build robust portfolios (HRP) that survive market crashes better than Markowitz models. | β Machine Learning |
17_optimal_execution_almgren.py |
Transaction Cost Analysis | Implements the Almgren-Chriss Model to calculate the optimal trading trajectory, balancing Volatility Risk vs. Market Impact cost. | β Quant Math |
| File Name | Type | Description | Status |
|---|---|---|---|
18_event_driven_backtester.py |
Architecture | A robust Event-Driven Engine handling Market, Signal, and Order events in a FIFO queue. Prevents look-ahead bias and simulates real exchange latency. | β Infrastructure |
19_triangular_arbitrage_graph.py |
Graph Theory | Uses NetworkX and the Bellman-Ford algorithm to detect negative cycles in Forex graphs, identifying risk-free Triangular Arbitrage opportunities. | β Math / Algo |
20_signal_processing_emd.py |
Physics | Demonstrates Signal Decomposition (similar to HHT/EMD) to separate high-frequency market noise from underlying structural trends using physics-based transforms. | β Signal Processing |
21_alternative_data_bert.py |
NLP / LLM | Utilizes FinBERT (Transformers) to analyze complex financial sentences, detecting nuanced sentiment that simple keyword algorithms miss. | β AI / Deep Learning |
| File Name | Type | Description | Status |
|---|---|---|---|
22_high_performance_numba.py |
Latency Eng | Uses JIT Compilation (Numba) to accelerate Python loops by 100x, achieving C++ speeds for High-Frequency Trading. | β Optimization |
23_market_making_avellaneda.py |
HFT Strategy | Implements the Avellaneda-Stoikov Model to dynamically adjust Bid-Ask spreads based on inventory risk. | β Market Making |
24_tail_risk_evt.py |
Risk Mgmt | Uses Extreme Value Theory (EVT) and Generalized Pareto Distributions to model "Black Swan" events beyond standard VaR. | β Fat-Tail Math |
25_covariance_shrinkage.py |
Quant Math | Applies Ledoit-Wolf Shrinkage to denoise Correlation Matrices, ensuring robust portfolio optimization even with limited data. | β Linear Algebra |
π΅οΈββοΈ Hidden Layer & Deep Research
| File Name | Type | Description | Status |
|---|---|---|---|
26_synthetic_data_gan.py |
Generative AI | Uses a GAN (Generative Adversarial Network) to generate "Deep Fake" market data. Solves the problem of limited historical data for training AI models. | β Research Grade |
27_copula_dependence.py |
Adv. Risk | Implements Copulas to model Tail Dependence. Captures "Crash Correlations" where assets move together only during extreme panic events. | β PhD Mathematics |
28_order_flow_imbalance.py |
Microstructure | Calculates Order Flow Imbalance (OFI) from Level 2 data. This is a primary predictive signal used by HFT firms to forecast the next tick. | β HFT Signal |
29_dynamic_hedge_kalman.py |
Adaptive Quant | Uses Kalman Filters to dynamically estimate the hedge ratio between two assets, allowing the strategy to adapt to changing market regimes instantly. | β Adaptive Algo |
| File Name | Type | Description | Status |
|---|---|---|---|
30_stochastic_volatility_heston.py |
Stochastic Calc | Simulates the Heston Model using Euler-Maruyama method. Models volatility as a dynamic process, capturing the "Leverage Effect" seen in real crashes. | β Grad-Level Math |
31_regime_detection_hmm.py |
Probabilistic | Uses Hidden Markov Models (HMM) to unsupervisedly classify market data into "Regimes" (Bull, Bear, Sideways) for regime-switching strategies. | β Machine Learning |
32_black_litterman_allocation.py |
Bayesian | Implements the Black-Litterman Model to blend "Market Equilibrium" with "Investor Views," solving the instability problems of standard Mean-Variance optimization. | β Asset Mgmt Standard |
33_kelly_criterion_sizing.py |
Risk Mgmt | Demonstrates the Kelly Criterion for optimal position sizing. Proves mathematically why over-leveraging a winning strategy leads to long-term ruin. | β Money Mgmt |
| File Name | Type | Description | Status |
|---|---|---|---|
34_fractional_diff_stationarity.py |
Financial ML | Implements Fractional Differentiation (Lopez de Prado). Transforms non-stationary price data into stationary features without erasing trend memory, enabling better AI training. | β ML Standard |
35_hawkes_process_arrival.py |
Microstructure | Simulates a Hawkes Process (Self-Exciting Point Process) to model trade arrival times. Captures the "viral" nature of liquidity and flash crashes. | β HFT Math |
36_defi_amm_simulation.py |
Crypto Quant | Simulates a Uniswap v2 AMM ( |
β Web3 |
37_satellite_vision_signal.py |
Alt Data | Demonstrates a Convolutional Neural Network (CNN) logic to extract "Car Counts" from satellite imagery, generating trading signals from alternative data sources. | β Computer Vision |
| File Name | Type | Description | Status |
|---|---|---|---|
38_yield_curve_nelson_siegel.py |
Fixed Income | Calibrates the Nelson-Siegel Model to government bond data to build the Yield Curve. Essential for pricing bonds and detecting interest rate opportunities. | β Rates Quant |
39_pca_risk_factors.py |
Risk Mgmt | Uses PCA (Principal Component Analysis) to decompose a portfolio into hidden Risk Factors (e.g., Market, Rates), revealing true diversification levels. | β Factor Model |
40_exotic_barrier_option.py |
Exotics | Prices a Knock-Out Barrier Option using Monte Carlo simulations. Demonstrates handling of "Path Dependent" derivatives where history matters. | β Structured Product |
41_pin_informed_trading.py |
Microstructure | Estimates the PIN (Probability of Informed Trading) metric. Detects toxic order flow by separating "Noise" volume from "Informed" volume. | β Market Micro |
| File Name | Type | Description | Status |
|---|---|---|---|
42_smart_order_router.py |
Execution Algo | A Smart Order Router (SOR) that splits orders across Fragmented Liquidity (NSE/BSE/Dark Pools) to minimize impact and fees. | β Institutional Algo |
43_market_data_parser_binary.py |
Low Level Eng | Demonstrates Binary Struct Parsing. Reads raw byte-streams (simulating ITCH/TBT protocols) directly, avoiding slow string parsing. | β HFT Engineering |
44_granger_causality_network.py |
Econometrics | Uses Granger Causality tests to determine "Lead-Lag" relationships between assets (e.g., Does Banking Sector lead IT Sector?). | β Macro Research |
45_quantum_annealing_opt.py |
Quantum Finance | Implements Simulated Annealing, a physics-based heuristic used in Quantum Computing (D-Wave) to solve NP-Hard portfolio optimization problems. | β Quantum Proxy |
| File Name | Type | Description | Status |
|---|---|---|---|
46_cuda_gpu_monte_carlo.py |
Supercomputing | Uses NVIDIA CUDA Kernels (via Numba) to run 10 million Monte Carlo simulations in parallel on the GPU. Reduces calculation time from minutes to milliseconds. | β Hardware Accel |
47_supply_chain_propagation.py |
Knowledge Graph | Models the global supply chain as a Directed Graph. Mathematically propagates shocks (e.g., Earthquake at TSMC) downstream to predict stock drops before news hits. | β Fundamental Quant |
48_clock_sync_ptp_sim.py |
HFT Infra | Simulates Precision Time Protocol (PTP). Synchronizes internal server clocks with exchange timestamps to the microsecond to prevent latency arbitrage. | β Low Latency |
49_limit_order_queue_estimator.py |
Microstructure | Tracks FIFO Queue Position. estimates exactly how many shares are ahead of your limit order and calculates "Time to Fill" based on trade velocity. | β Execution Edge |
| File Name | Type | Description | Status |
|---|---|---|---|
50_trade_surveillance_wash_graph.py |
RegTech | Uses Graph Cycle Detection to identify circular "Wash Trading" rings, a critical compliance requirement for all major exchanges. | β Compliance |
51_weather_derivative_pricing.py |
Exotic Asset | Prices HDD/CDD Weather Options using temperature simulation. Used by energy desks to hedge against "Warm Winters." | β Commodities |
52_tax_loss_harvesting_algo.py |
Tax Alpha | Automates Tax Loss Harvesting: realizing losses to offset gains while maintaining market exposure via correlated substitutes (Spider Strategy). | β WealthTech |
53_distributed_grid_ray.py |
Cloud Infra | Implements Distributed Computing using Ray. Demonstrates how to scale analytics from a single laptop to a massive server cluster. |
β Scalability |
To run these scripts, you need Python installed along with the following financial libraries:
pip install numpy pandas matplotlib scipy yfinance statsmodels scikit-learn network textblob nselib tensorflow torch transformers xgboost lightgbm hmmlearn numba ray simpy
### β οΈ Legal Disclaimer & Risk Disclosure
**IMPORTANT: PLEASE READ BEFORE USING THIS REPOSITORY.**
This repository contains advanced financial algorithms, including High-Frequency Trading (HFT) logic, Derivatives pricing models, and Tax optimization strategies. By accessing or using this code, you agree to the following terms:
### 1. No Financial Advice (SEBI/SEC Disclaimer)
* **Educational Use Only:** The code provided here is strictly for **Research and Educational purposes**. It demonstrates the *mathematics* and *engineering* behind Quantitative Finance.
* **Not a Recommendation:** Nothing in this repository constitutes investment advice, financial promotion, or a recommendation to buy/sell any security.
* **No Fiduciary Duty:** The author is not a SEBI Registered Investment Advisor (RIA), Portfolio Manager, or Broker-Dealer. Consult a certified financial professional before making investment decisions.
### 2. Algorithmic Trading Risks
* **Real Money Warning:** Using Python scripts to trade real money carries significant risk. A software bug, API failure, or internet outage can result in **100% loss of capital** in seconds.
* **No Warranty:** This software is provided "AS IS", without warranty of any kind. The author is not liable for any financial losses incurred from running these scripts in a live environment (Paper Trading or Real Trading).
* **Past Performance:** Backtesting results (e.g., File 01, 05) are simulated on historical data. **Past performance is not indicative of future results.** Real-world execution involves slippage, liquidity constraints, and transaction costs that may not be fully modeled here.
### 3. Compliance & Market Manipulation
* **Market Integrity:** Modules like `50_trade_surveillance_wash_graph.py` demonstrate how to **detect** illegal activity (Wash Trading). Using code to **execute** Wash Trades, Spoofing, or Layering is a criminal offense under SEBI (India) and SEC (USA) regulations.
* **HFT Regulations:** High-Frequency Trading (File 46, 48) is highly regulated. Deploying HFT strategies often requires specific exchange approvals, co-location agreements, and stress-testing certifications.
### 4. Technical & Data Risks
* **Data Accuracy:** Data fetched via open-source libraries (`yfinance`, `nselib`) is not real-time tick data and may contain errors or delays. Do not rely on it for precision pricing.
* **Hardware Usage:** Scripts involving GPU acceleration (`46_cuda_gpu_monte_carlo.py`) or Distributed Computing (`53_distributed_grid_ray.py`) can cause high hardware load. The author is not responsible for hardware damage or cloud computing costs.
### 5. Tax & Legal
* **Not Tax Advice:** The Tax Loss Harvesting algorithm (`52_tax_loss_harvesting_algo.py`) is a simulation of logic. Tax laws (Income Tax Act, 1961) change frequently. Do not use this code for tax filing without verifying with a Chartered Accountant.
---
### π License
This project is licensed under the **MIT License** - you are free to use, modify, and distribute this software, but you hold the author harmless from any liability.