How to Learn Supply Chain Analytics in 2027 with Claude

A playbook to learn supply chain analytics for professionals without operational experience using 5-minute case studies and a coding agent.

Need Help?
Subscribed! Error
How to Learn Supply Chain Analytics in 2027 with Claude

When I started selling analytics solutions to multinational companies, I needed AI engineers and data scientists to deploy them with me.

Because supply chain processes are complex, I searched for profiles who could capture operational teams' needs and translate them into a product.

I quickly realised that what I was looking for barely existed.

Supply chain professionals know the processes, but they mostly work in Excel.

The overlap I could not find on the market (Image by Samir Saci)

And the data scientists I met usually built their careers in finance or marketing, not in warehouses and factories.

So I stopped looking for a unicorn and designed a method to teach supply chain analytics to professionals who have only half the knowledge.

Because coding has become a commodity, analytics professionals now need a domain expertise.

Coding agents are game changers for people willing to learn supply chain analytics:

  • Data scientists: as Claude now writes the code you were hired for, use operational case studies to learn how processes work
  • Supply chain professionals: you can ask the agent to adapt the code from each case study, so you don't need to learn programming to build a tool.

This is a new way for anyone to enter this domain, and it needs analytics professionals with the keys to understand complex operations.

Whether you are a data scientist or a supply chain professional who can’t code, this article shows you how to upskill faster

  • How I structure my case studies to teach both worlds at once
  • How coding agents help learners practise without guidance

I will illustrate my methodology with a real case study of production planning in a factory.

You can practise with the source code shared at the end of the article.

Table of Contents
I. You need to accumulate business and operational experience
II. How I teach a supply chain concept in 5 minutes
III. How Claude Code accelerates learning
IV. Learn with case studies, not with textbooks
Discover 22+ case studies shared on my YouTube channel, Supply Science

You need to accumulate business and operational experience

I strongly believe now that analytics professionals need domain expertise to stay relevant, and well paid, in the current market.

Many large companies that invested in AI projects, because of the FOMO created by the hype, are now stuck in implementation.

I will illustrate my point with a project at a manufacturing company for which I needed two AI engineers to deploy agents that would support production planners.

My main recruiting issue wasn't technical.

I couldn't find analytics professionals who had put a foot in a factory or ever talked to a production planner.

Why does it matter?

This operational understanding, and the ability to translate an operational problem into a solution, represent about 80% of our projects' success.

Customers will not tell you that they need a time series forecasting model using exponential smoothing coupled with a periodic review policy.

They will tell you that they have an inventory problem.

Nobody asks for exponential smoothing (Image by Samir Saci)

Between these two sentences, someone has to ask:

  • Which items are concerned?
  • Which decisions are made with the stock levels?
  • Which data is available to support decisions?

This isn't a coding skill, and a coding agent will not do it for you.

Unfortunately, this experience usually takes years of operational roles to accumulate.

This competitive advantage helped me start my consulting business, but it quickly became a bottleneck to our growth.

I could not wait years to build my team of consultants.

So I packaged the experience I'd gained in 10 years into case studies tailored to teach these concepts in 5-minute lessons, so I could upskill my team fast.

Let me show you how I structure them.

How to teach a supply chain concept in 5 minutes

The five minutes refer to what I need to spend to film myself explaining the case study. 

After this, I can outsource the explanations to a Claude skill.

Let’s imagine a case study designed to teach the basics of production planning optimisation to:

  • A data scientist who does not know what a production setup is
  • A supply chain professional who has never written an optimisation algorithm
One case study, two audiences (Image by Samir Saci)

Each lesson follows the same three parts, and each part teaches something different to each of them:

  • The context helps you link an operational situation and its translation into data
  • A problem statement is formulated using operational wording to upskill analytics professionals
  • The solution is documented to help supply chain professionals understand each line of code.

The goal is to help any professional acquire the minimal skills to start working on the case study.

Let’s test this approach with the production planning optimisation project.

Context: start with the operations, not with the model

We imagine you are the production planning manager of a small factory producing radio equipment for local and international markets.

Customers send purchase orders to your commercial team with quantities and expected delivery dates.

From the purchase order to the factory (Image by Samir Saci)

For example:

  • Customer A ordered 100 units of SKU XX for November 23rd
  • Customer B orders 150 units of SKU YY for November 30th

Your role is to schedule production to deliver on time with the minimum total cost of production:

  • Setup costs: fixed costs you pay each time you set up a production line
  • Production costs: variable costs per unit produced
  • Holding costs: the cost of storing one unit for one month

In our example, customers ordered products for the next 12 months.

12 months of purchase orders (Image by Samir Saci)

For the data scientist, this part is the real lesson: what a purchase order, a setup and inventory on hand mean on the shop floor.

For our supply chain professional, it is her daily job, written as a data table.

Problem statement: name the trade-off in operational terms

Your boss, the planning manager, has two objectives that pull in opposite directions.

We need to minimise the inventory, but I don’t want too many production batches.
Two objectives, one plan (Image by Samir Saci)

To illustrate, imagine a Formula 1 race with cars that are slower when their fuel tank is full but lose precious seconds each time they pit.

If you follow the Chief Financial Officer, who wants to optimise cash flow, you will produce exactly what you need each month.

You will have no excess inventory (cash sleeping in your warehouse), but you pay a setup every month.

Plan 1: produce every month (Image by Samir Saci)

On the other hand, if you follow the factory manager, you will produce large batches to reduce setups.

Unfortunately, this builds large stock, which increases holding costs and traps cash in sleeping inventory.

Plan 2: build stock to save setups (Image by Samir Saci)

Between these two extremes, you struggle to find an optimal balance.

There is an algorithm that balances the two automatically.

This is where the two audiences meet:

  • The data scientist understands why an optimisation is needed for an operational reason, not because the method is elegant.
  • The supply chain professional sees the trade-off they manage every day written as a cost to minimise.

They can now provide the same output with the help of a coding agent (and limited guidance from my side) using the code presented in the solution part.

Solution: one method, the code, and the result

This is the delicate part for the supply chain professionals as we start to talk about algorithms and lines of code.

In this case study, I quickly introduce the Wagner-Whitin algorithm.

Wagner-Whitin Algorithm explained in one visual — (Image by Samir Saci)

I explain in simple terms how the algorithm works, without too many details, because I know neither audience will ever build it from scratch.

My tip for teaching a non-tech audience is to use an actual example and do the optimisation by hand.

Visual explaination of the algorithm — (Image by Samir Saci)

I used this table to explain, step by step, how the algorithm works without showing a single line of code.

I want to avoid losing my supply chain professionals with a dynamic programming algorithm they will never write from scratch.

I solely focus on what is important:

  • Inputs: it takes a setup cost of 500 $, a holding cost of 1 $/unit/month and a production cost of 50 $ per unit
  • Outputs: the algorithm finds a plan with four production runs (Month 1: produce 550 units for the first 5 months, Month 6: produce 450 units for months 6, 7 and 8 …)
Four production runs, and an inventory that follows the demand (Image by Samir Saci)

My goal is to avoid the friction that an overly complex explanation can create for someone trying to learn the tool.

If someone wants to dig into the code, she can ask the coding agent.

Holding and setup costs, balanced over 12 months (Image by Samir Saci)

The key insight: this cumulative cost chart shows how the algorithm balances the two objectives over the year.

In 5 minutes, both audiences have seen how a piece of code solved an operational problem, and now they can learn how to use it.

With the source code and a sample dataset, anyone can reproduce the result and answer my questions:

  • Q1: How many setups would we have with a holding cost of $ 2?
  • Q2: Which scenario is suitable if we have a 1-month delivery lead time?

The first question is challenging for supply chain professionals, while the second one can’t be answered by analytics professionals.

Hopefully, they have Claude Code to support them.

How Claude Code accelerates learning

Unfortunately, I can’t spend hours training my freelance consultants in supply chain analytics.

So I created a Claude Code skill that guides them through the process, so they can practise on their own with tricky questions like the one I shared in the previous section.

A learning process assisted by Claude (Image by Samir Saci)

The process has four steps:

  1. They load the skill and give Claude the title of the video
  2. Claude pulls the source code and a sample dataset
  3. The skill deploys it in a local app with all the input parameters to play
  4. I send them requests that simulate user interaction, and Claude explains the changes in the code

The third step changes everything for the supply chain professionals who have never run a Python script.

But also for analytics professionals who may not always know how to translate a business question into code change.

The case study, turned into an app they can play with (Image by Samir Saci)

The case study becomes an app with parameters they can change, not a notebook they have to read.

The fourth step is where they really learn:

  • Q1: How many setups would we have with a holding cost of $ 2?
    The coding agent explains that no code change is required, just switching the holding cost parameter and re-running.
  • Q2: Which scenario is suitable if we have a 1-month delivery lead time?
    The coding agent explains what a 1-month delivery lead time is and how it affects the algorithm, adapts the code, and explains the results.

Because Claude or GPT explains each change, they learn the analytics concepts when they need them.

It is beyond learning how to write code, its more how to question it.

Learn with case studies, not with textbooks

In 2027, the value of analytics professionals will reside more in translating business problems into analytics solutions than in pure coding skills.

You no longer need to spend hours reading academic books (like I did) to learn how to use analytics and AI to solve operational problems.

Two ways to learn supply chain analytics (Image by Samir Saci)

Now you just need someone to introduce a real case study in a concise format and practise.

If you want to start, I have shared 22+ case studies in a playlist of 5-minute videos: each one presents a real operational problem, the method that solved it, and the source code with a sample dataset.

The production planning case of this article (with the source code) is one of them:

About me

I am Samir. I have ten years of supply chain experience across Asia and Europe, and I build the analytics and AI tools at LogiGreen.

I record what I learn on Supply Science, and the rest is on my blog.

Let’s connect on LinkedIn and Twitter.

Need Help?