.

A circular economy is a system where waste is minimized and resources are continuously reused or recycled.

As the current linear economic model reaches its limits, discussions of new circular business models become increasingly prominent.
What is holding us back?
These discussions mainly focus on
- The operational and business obstacles blocking the transition
- Alternative strategies to increase the use of recycled materials
- Rental models to reduce the environmental footprint
As a data science manager of a retail company, how can you support this transition?
We can leverage the data generated by systems to overcome these barriers by identifying opportunities to create a sustainable circular economy with Data Science.

In this article, we will assume the role of a data science manager who has been asked to support the operational transformation of a fashion retail company.
Summary
I. Transition to a Circular Economy
1. What is the environmental impact of a T-shirt?
2. Data-driven Process Design
II. Overcoming the Operational Challenges
1. The Opacity of Supply Chain Networks
2. The Low Residual Value of Used Products
III. Material Efficiency & Recycled Materials Usage
1. Raw Material Optimization with Linear Programming
2. Supply Chain Network Optimization
IV. Conclusion
Transition to a Circular Economy
The evolution from a linear model to a circular economy is an ongoing process with significant business and operational implications.
This shift is not just about waste management or recycling.
It requires a holistic change in how we design, produce, sell and use goods or services.

Before implementing a circular economy, the first step is to estimate the environmental impact of our current linear model.
What is the environmental impact of a T-shirt?
Let’s take the example of a T-shirt you bought in a fast-fashion store.
What is its environmental impact along its life cycle?
Life cycle assessment (LCA) is a methodology for evaluating the environmental impacts of a product or service over its entire life cycle.

- Raw materials are sourced from different suppliers that are using natural resources and energy.
- Manufacturing sites transform these materials into finished products using natural resources while emitting pollutants and CO2
- Finished products are delivered to stores and sold to final customers
- Customers are using the products until disposal
How can we support for the automation of Life Cycle Assessment?
This descriptive analytics methodology can be automated using Business Intelligence solutions implemented by our analytics team.
The challenge is to collect and process transactional data
-
From different systems that may not communicate with each other Factory Management Systems vs. Warehouse Management Systems
-
With different formats (Unstructured vs. Structured) Excel Utility Usage Reports vs. WMS Transactional Database(s)

💡 Data analysts and engineers can implement pipelines using a central data warehouse to collect and process raw data to feed LCA calculations.

The final result can be a self-service database of harmonized tables containing transactional records covering the full cycle from raw material collection to store delivery.
💡 For more details,
Your sustainability department can then use these tables to run calculations and estimate each process’s resource usage or CO2 emissions.
What can we do to reduce our environmental footprint?
Data Analytics for Solution Design
Now that you have automated the Life Cycle Assessment, your sustainability team has been able to set the baseline.
The total CO2 emissions for 2022 are 75k Tons Co2eq.
Following the United Nations Sustainable Development Goals (SDG), your company committed to a 30% reduction by 2023.
The next step is to build a roadmap to reach this target.
I have previously shared data-driven methodologies to implement decarbonization initiatives.
- Sustainable Sourcing: select the set of suppliers that minimizes the environmental impact of your raw materials sourcing
- Sustainable Supply Chain Optimization: design an optimal network of factories and warehouses to minimize the emissions
- Circular Economy: create a logistic chain to collect and reuse returned items from customers to save raw materials
A circular model can reduce the highest carbon emissions as it directly impacts the product.
However, the case studies above mainly focused on generating insightful prescriptions using advanced analytics.
Now that your model told us what to do. How can you support the implementation?
Because such a transition can completely disrupt your current supply chain operations.
Logistics Team: How do we organize our truck fleet to collect returned items?
Logistics operations expect support to ensure a smooth implementation and avoid disrupting the business or impacting profitability.
The next section will show how data science can support this operational transformation.
Overcoming the Operational Challenges
While the shift to a circular economy looks promising, it comes with various operational challenges.
The Opacity of Supply Chain Networks
This is the primary barrier to the transition towards a circular economy.
From which factory this batch of finished goods is coming from?
In traditional linear economies, the origin and journey of goods from raw materials to the final product often need to be clarified.

Your company might not understand its Supply Chain clearly beyond their immediate suppliers and customers.
- Can you track the production facility of any item sold in your stores?
- Can you link a finished product leaving the factory with the batch of raw materials used to manufacture it?
A lack of transparency makes tracing products back to their source difficult.
This is creating a stumbling block in adopting circular economic practices.

Understanding a product’s lifecycle (from raw materials to disposal) is crucial for implementing efficient recycling and reusing strategies in a circular economy.
With opacity, we cannot ensure that materials are sourced sustainably, used efficiently and recycled properly.
You might also miss opportunities to reduce waste, streamline operations and use resources more efficiently.

An optimal circular economy would require a minimum set of performance indicators like
-
Tracking of production and transportation KPIs with Production Adherence (%), Replenishment Lead Time (Days)
-
Measuring the performance of reverse logistics with Logistic Costs (Euros/piece), Ratio of Returned Items (%) and Collection Lead Time (Days)
- Identifying potential improvements for the recycling process with Recovery Rate (%), Contamination Stream (%) and Processing Lead Time (Days)
How can we monitor these KPIs using analytics solutions?
These logistics and manufacturing KPIs require data from multiple systems with different data formats and database structures.

💡 Your team can play a crucial role in addressing these issues by
- Connecting to systems that track your products along the value chain and gathering data with timestamps
- Store and process this data to create a central source of information that can be used to create reports, dashboards and optimization models
- Implement automated reporting tools with KPIs designed by Supply Chain and Sustainability teams
This comprehensive view identifies inefficiencies, provides traceability to customers and facilitates the transition.
For more analytics solutions for supply chain traceability,
Logistic Performance Management Using Data Analytics
Can we ensure the economical viability to this model using data?
The Low Residual Value of Used Products
In our existing linear economy, products are designed for consumption and disposal but not for reuse or recycling.
Product residual value refers to the remaining worth of a product after it has been used and completed its initial lifecycle.
Once used, these products often have little residual value.
How can we design a profitable circular model?
Therefore, the costs associated with collecting used products for recycling often outweigh the value of the materials recovered.
This endangers the economic viability of circular business models and discourages businesses from transitioning.

For instance, if we take the example of our T-shirt
- Forward Logistics is cost-efficient as we are delivering t-shirts by full containers with large trucks using optimized routing
- Reverse Logistics, in comparison, is extremely expensive as we collect used garments by piece with complex sorting and recycling flows that require customized processes
As reverse logistics operations can become highly complex, it’s easy to face the situation that recycling becomes more expensive than disposal.
Data analytics can play a crucial role in addressing these issues.
We can simulate these circular models to find the optimal setup to bring efficiency and cost-effectiveness.
- Streamline reverse logistics operations using optimization models to minimize the cost of collection and sorting of used items
- Designing alternative business models like subscription models for which items are rented instead of sold
Data Science for Sustainability – Simulate a Circular Economy
With these additional simulation models, you can support the implementation of a profitable reverse flow to recycle (or reuse) your products.
Material Efficiency & Recycled Materials Usage
Material efficiency becomes a predominant concern as we move towards a circular model.
Making products with fewer materials and minimizing waste in production processes can bring substantial economic and environmental benefits.
How to produce to ensure ensure the longevity and reusability of finished goods?
Raw Material Optimization with Linear Programming
Efficient use of materials can drastically reduce waste and support sustainable use of resources.
Different types of fabrics are at our disposal, including cotton, polyester, linen and silk.

Each type of fabric has different costs and attributes, such as durability, comfort and environmental impact.
The manufacturer’s goal is to minimize the overall cost of production while meeting the necessary quality and sustainability standards.

What is the best mix to meet our profitability and sustainability goals?
This is a multi-dimensional optimization problem where we are trying to optimize for cost and sustainability under certain constraints.
Example of T-shirt manufacturingLet us imagine a scenario in which the T-shirt must contain
- At least 40% cotton for comfort
- Not more than 30% polyester due to sustainability guidelines
- Silk must not exceed 10% of the total material
This problem can be modelled and solved using linear (or non-linear) optimization with Python.

- Parameters: The quantity of each raw material used to produce a T-shirt
- Constraints: the one listed above
- Objective function: minimize environmental footprint, minimize the cost or a mix of both
Your team can use libraries like PuLP or SciPy to create an optimization model for testing several objective functions and eventually find the perfect mix of materials.
Have a look at this example for more details,
Raw Materials Optimization for Food Manufacturing with Python
Where do we need to produce to deliver our markets most sustainably?
Supply Chain Network Optimization
To introduce reverse flow processes for recycling, we have to redefine our supply chain network completely.
Supply chain optimization can help us to make the best use of data analytics to find the optimal combination of factories, distribution and recycling centres that minimize cost impacts.

A linear programming model with Python can help us by
- Selecting the right locations for our recycling centres
- Optimizing the flows of used items collection
- Sizing the capacity of sorting and recycling centres
The objective is to minimize the cost of collecting, sorting and recycling used items for a profitable and sustainable circular model.
For more details on how to create a network optimization model,
I hope these examples gave you enough inspiration to support your sustainability department in its efforts to transition to a more viable economic model.
Conclusion
Data science can be a powerful enabler in transitioning towards a profitable circular economy by overcoming complex barriers and optimising resources.
In the future, analytics teams will likely play a key role in helping companies transition to circular models.
How can you share these insights to operational teams?
Deploy your tools on web applications.
Providing easy access to everyone in the organization is a great way to support the implementation of a data-driven prescriptive model.
You can productize this solution by deploying it on a web application that operational and business people can use.
I have deployed three models, which I have presented in my articles, using the VIKTOR platform.
Sustainable Supply Chain Optimization Web App
![Access the Application to try it! - [App]](https://towardsdatascience.com/wp-content/uploads/2023/10/0wrrOXkNSpIP7tr6I.png)
ABC Analysis and Pareto Chart Application
![Access the Application to try it! - [App]](https://towardsdatascience.com/wp-content/uploads/2023/06/1hh2WAM-xs3eAjwPrnebatw.png)
Production Planning
![Access the Application to try it! - [App]](https://towardsdatascience.com/wp-content/uploads/2023/10/1iLp1e8Xi_pwxtLvtuNeHug.png)
For more details on how I did, check this article
About Me
Let’s connect on Linkedin and Twitter. I am a Supply Chain Engineer who uses data analytics to improve logistics operations and reduce costs.
For consulting or advice on analytics and sustainable supply chain transformation, feel free to contact me via Logigreen Consulting.
If you are interested in Data Analytics and Supply Chain, look at my website.
💌 New articles straight in your inbox for free: Newsletter 📘 Your complete guide for Supply Chain Analytics: Analytics Cheat Sheet