Actura | Actuarial Valuation & Risk Platform

📌 Overview

Actura is an enterprise-grade actuarial platform that bridges the gap between complex risk mathematics and modern UI/UX. It provides actuaries and risk engineers with a powerful Python-based valuation engine (handling everything from Monte Carlo simulations to IFRS 17 trajectories) paired with a sleek, neo-skeuomorphic Vue 3 dashboard.

📸 Screenshots

Valuation Dashboard

Real-time stochastic projection, tail risk (CVaR) analysis, and comprehensive reserve profiles.

Contract Logic Blueprint Builder

A visual, node-based DAG (Directed Acyclic Graph) editor for designing insurance cash flow blueprints without writing code.

---

✨ Features

• 🧩 Visual Logic Builder: Drag-and-drop DAG editor to map out premiums, benefits, and decrement models seamlessly.
• 📈 Stochastic Engine: Integrated Economic Scenario Generators (ESG) and advanced Monte Carlo simulations.
• 📊 Risk Analytics: Built-in calculation for BEL, VaR 95%, CVaR / CTE 95, and quantile fan-charts.
• ⚙️ IFRS 17 Ready: Automated CSM (Contractual Service Margin) mechanics and Risk Adjustment tracking.
• ⚡ High-Performance Backend: Powered by FastAPI for rapid deterministic and stochastic valuations.
• 🐳 Docker-Ready: Fully containerized architecture with Nginx reverse proxy for immediate, production-like deployment.

🏛️ Architecture

Actura uses a modern decoupled architecture:

1. Frontend: Vue 3 SPA handling graph interactions (Vue Flow) and charting (ECharts).
2. Backend: FastAPI providing REST endpoints to handle math-heavy actuarial modeling, state management, and DB interactions.
3. Database: PostgreSQL (via SQLAlchemy) to store projects, blueprints, and stochastic results.
4. Proxy: Nginx for static serving and routing.

See Architecture Documentation for deeper details.

⚙️ How it works

Actura evaluates insurance cash flows as a Directed Acyclic Graph (DAG).

• Nodes represent financial events (e.g. premium collection, mortality decrements, expense payouts).
• Edges dictate the flow of policyholder states and cash across the timeline. The graph is compiled by the Vue 3 frontend, saved as a JSON blueprint to the database, and executed by the Python actuarial engine over projection timesteps.

See Blueprint Engine for more details.

🔄 Example workflow

1. Project Setup: Create a new project and select an actuarial preset (e.g., 20-Year Term Life).
2. Blueprint Design: In the Visual Builder, drag and drop nodes to define logic (e.g., attach an Expense node to the Inflow).
3. Simulation: Hit "Run Simulation". The backend parses the DAG and generates cash flow trajectories.
4. Risk Dashboard: View the aggregated results (BEL, Net Cash Flow, Stochastic Fan Charts) on the main dashboard.

🛠️ Tech stack

• Backend: Python 3.10+, FastAPI, SQLAlchemy, NumPy, Pandas, SciPy.
• Frontend: Vue 3 (Composition API), Vite, Tailwind CSS, Vue Flow, Apache ECharts.
• Infrastructure: Docker, Docker Compose, Nginx, PostgreSQL.

🚀 Installation

Quick Start (Docker)

1. Clone the repository

   git clone https://github.com/rytsia1/actuarial-valuation-engine.git
   cd actuarial-valuation-engine

2. Build and spin up the containers

   docker compose up --build -d

3. Access the Platform
   • 🖥️ UI Dashboard: http://localhost:3000
   • ⚙️ API Docs: http://localhost:8000/docs

For local development setup without Docker, check out Development Guide.

📡 API

The backend exposes a REST API via FastAPI.

• /projects/: CRUD operations for workspaces.
• /projects/{id}/blueprints/: Blueprint DAG persistence.
• /projects/{id}/valuations/: Trigger stochastic or deterministic engine runs.

See the full API Documentation.

🧪 Testing

Actura leverages pytest for robust numerical validation.

docker exec -it actura_postgres pytest tests/

We assert calculations against standard actuarial tables (e.g., SOA ILT) to guarantee numerical precision.

⚡ Performance

• Vectorized array operations via NumPy allows for 1000+ stochastic scenarios to be run in under a second.
• Vue Flow handles 100+ DAG nodes efficiently using hardware-accelerated rendering.
• ECharts leverages Canvas/WebGL for smooth 60fps charting of large stochastic datasets.

🗺️ Roadmap

• Export to Excel/CSV for IFRS 17 disclosures
• Collaborative real-time blueprint editing
• Integration with external ESG providers (e.g., Moody's / Barrie & Hibbert)
• Additional decrement tables support (Morbidity, Multi-state transitions)