Skip to content
Shyenova

Programme 03

Analytics for Supply Chain Professionals

By Friday you will have built a forecaster, an inventory optimiser, a network model and a simulation, and they will run on your own data.
Level
Intermediate
Duration
5 days
Cohort
10–16 players
Formats
In-person, Virtual live

There is a persistent and expensive gap between supply chain practitioners who understand the problems and data scientists who can build the models. Practitioners commission analytics they cannot evaluate; data scientists build technically sound models that answer the wrong question.

This programme closes the gap from the practitioner side. It is not a general data science course with supply chain examples bolted on, every session builds a working model of a real supply chain problem, in Python, on realistic data. Players leave with a repository of code they wrote and understand.

Learning outcomes

What a player can do afterwards.

Written as capabilities rather than topics covered, because the second is easy to deliver and impossible to verify.
  1. 01Build and validate demand forecasts, and measure accuracy in ways that reflect the decision
  2. 02Compute multi-echelon safety stock and produce a service-versus-inventory curve
  3. 03Formulate and solve a network design problem as a mixed-integer program
  4. 04Build a discrete-event simulation of a flow and use it to test policy
  5. 05Judge whether an analytical result is sound and where it will break

Curriculum

5 days, hour by hour.

Published in full because you should be able to judge the substance before committing a team's week to it.
Day 1

Data foundations

Anchor exercise

Build a clean, validated analytical dataset from raw transactional extracts.

How we apply this: Data Foundations
  • Python and pandas for supply chain data structures
  • Transactional data: shipments, orders, inventory positions
  • Time series construction, aggregation and hierarchy
  • Data quality diagnosis: the failure modes that invalidate everything downstream
  • Visual exploration and the plots that actually reveal structure
Day 2

Forecasting

Anchor exercise

Build a segmented forecasting pipeline with automatic model selection and honest backtesting.

How we apply this: Forecasting
  • Baseline methods and why naive forecasts are the benchmark that matters
  • Exponential smoothing and state space models
  • Handling seasonality, promotions and calendar effects
  • Intermittent demand: Croston, SBA and the methods built for sparse data
  • Machine learning for demand: where it genuinely wins and where it does not
  • Backtesting properly, and the accuracy metric appropriate to the decision
Day 3

Inventory optimisation

Anchor exercise

Build a multi-echelon inventory optimiser and generate the efficient frontier for a supplied network.

How we apply this: Demand & Inventory Planning
  • Newsvendor and the critical fractile, derived and implemented
  • Safety stock under demand and lead-time variability
  • Reorder point, order-up-to and periodic review policies
  • Multi-echelon optimisation and the guaranteed-service model
  • Building a service-versus-inventory efficient frontier
Day 4

Network and optimisation

Anchor exercise

Formulate and solve a facility location problem end to end, then run scenario sensitivity.

How we apply this: Decision Science
  • Linear and mixed-integer programming for supply chain problems
  • Facility location formulation and solution
  • Transportation and assignment problems
  • Capacity and sourcing allocation models
  • Sensitivity analysis and reading a dual value
  • Recognising when an optimisation model is lying to you
Day 5

Simulation and delivery

Anchor exercise

Build a simulation of a replenishment policy under stochastic demand and lead time, and use it to select a policy.

How we apply this: Digital Twin
  • Discrete-event simulation for supply chain flows
  • Monte Carlo for risk and variability propagation
  • Testing policy under uncertainty rather than at the mean
  • Communicating analytical results to decision makers
  • Productionising: from notebook to something a colleague can run
  • Applied session on players' own problems

Assessment

How it is marked.

Players complete daily builds that are reviewed for correctness, and a capstone applied to a problem from their own organisation. The capstone is presented on the final afternoon and assessed on analytical soundness and on the clarity of the recommendation to a non-technical decision maker.

Questions

The ones we get asked.

How much Python do players need beforehand?

Enough to be dangerous rather than fluent: variables, loops, functions, and some exposure to pandas. We send a preparatory module roughly four hours long that brings a competent spreadsheet user to the required starting point, and we check completion before the programme runs. Absolute beginners will struggle from day two onward.

Can players use their own data?

For the capstone, yes, and we encourage it. Days one to four use prepared datasets so the whole cohort progresses together, but the Friday capstone is deliberately applied to a real problem the player owns. Data protection arrangements are agreed with the sponsoring organisation beforehand.

Is this a data science course?

No. A data science course teaches methods and uses supply chain data as an example. This teaches supply chain problems and uses whichever method the problem requires, which is why a mixed-integer program and a discrete-event simulation sit alongside the forecasting content rather than being treated as separate disciplines.

What do players take away?

A documented repository containing every model built during the week, the datasets, and their capstone. All of it is theirs to reuse. Several past cohorts have taken the day three inventory optimiser directly into production with modest adaptation.

Should players attend Fundamentals first?

If they have not worked inside a supply chain, yes. The models built here assume the player already knows what a safety stock is protecting against and why a network decision is a trade-off rather than an optimisation. For a planner or analyst with a few years in the function, this programme stands alone.

Enquire

Run Analytics for your team.

Tell us the roles, the current level and what the team has to do differently afterwards. We will confirm whether this is the right programme or point you at a better fit.

Or write directly to hello@shyenova.com