Waste Networks Analysis

A comprehensive library for analyzing waste in supply chain networks using advanced network analysis and causal inference.

What is This?

Think of a food supply chain as a complex web of relationships between different players: farmers, processors, warehouses, stores, and service providers. Each connection in this web can lead to food waste, but understanding exactly how and why waste occurs is challenging. This tool helps solve that challenge by:

1. Mapping the Network: Creating a digital twin of your supply chain that shows how food, services, and money flow between different players
2. Finding Waste Hotspots: Using advanced math to identify where waste is most likely to occur
3. Understanding Causes: Using AI and statistics to figure out what factors (like storage time or temperature) most affect waste
4. Testing Solutions: Allowing you to simulate different solutions (like better cold storage) to see their impact

Key Concepts

Players in the Network

• Initial Producers (e.g., farms): Where food enters the system
• Food Processors (e.g., packaging facilities): Where food is transformed
• Food Handlers (e.g., warehouses): Where food is stored and moved
• End Consumers (e.g., stores): Where food exits the system
• Service Providers (e.g., cold chain services): Who help reduce waste

Types of Connections

• Inventory Flow (solid blue lines): Shows how food moves
• Service Flow (dashed red lines): Shows who's helping who
• Currency Flow (dotted green lines): Shows how money moves

Waste Calculation

We calculate waste in three ways:

1. Static: Fixed percentage (e.g., 5% always lost)
2. Time-based: Increases with time (e.g., 1% per day)
3. Multi-factor: Based on conditions (e.g., temperature, humidity)

Features

• Advanced network model with flow-specific path finding
• Bayesian causal analysis with regression capabilities
• Visualization tools for network structure and statistical analysis
• Inventory loss prediction and analysis

Installation

1. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

Understanding the Output

1. Network Map (network_visualization.png)

This is like a Google Maps for your supply chain:

• Each dot (node) is a player in your system
• Lines between dots show how they're connected
• Colors tell you what type of player each dot represents
• Line styles show what's flowing between players

Key things to look for:

• Thicker lines = more flow
• Red highlights = potential problem areas
• Numbers on lines = amount of waste

2. Waste Analysis (waste_breakdown.png)

This chart shows where waste is happening:

• Each bar represents a location
• Height of bar shows how much waste occurs there
• Colors match the network map for easy reference
• Total waste shown at top

3. Causal Analysis (causal_results.png)

This helps understand why waste occurs:

• Numbers show how strong each factor's effect is
• Larger numbers = stronger effect
• ± shows uncertainty in the measurement
• R² score (0-1) shows how well we understand the system

Example interpretation:

Storage Time: 0.32 ± 0.017 (R² = 0.85)

Means:

• Every day in storage increases waste by about 0.32%
• We're quite certain (small ± number)
• Our model explains 85% of waste variation (good fit)

Examples

See the examples/ directory for detailed examples:

1. Basic Network (basic_example.py)

Shows how to:

• Create a simple supply chain
• Add basic waste calculations
• Visualize the network

2. Advanced Analysis (advanced_example.py)

Shows how to:

• Use causal analysis
• Model complex relationships
• Test different solutions

3. Real-world Demo (real_world_demo.py)

A complete example using:

• Real-world-like data
• Multiple node types
• All flow types
• Complex waste functions

How to Use Results

1. Find Problem Areas

   • Look for red highlights in network map
   • Check highest bars in waste breakdown
   • Focus on strongest effects in causal analysis

2. Choose Solutions

   • Use service providers where effects are strongest
   • Optimize paths with least waste
   • Target interventions based on causal factors

3. Monitor Progress

   • Track total system waste over time
   • Watch for changes in causal factors
   • Measure solution effectiveness

Documentation

For detailed mathematical formulation and methodology, see the whitepaper in the whitepaper/ directory.

Contributing

Contributions are welcome! Please read our contributing guidelines and code of conduct.

License

This project is licensed under the MIT License - see the LICENSE file for details.