Supply Chain and Forecasting with Python
Supply Chain and Forecasting is an intermediate-level course of 5 lessons (about 1 h 24 min), with 15 exercises and 15 quiz questions. Prerequisites: Business Metrics and KPIs. Exercises run in the browser, with nothing to install.
- Business analytics
- Intermediate
- 5 lessons
- 1 h 24 min
- 15 exercises
- Exercises run in the browser
The course in brief
This course applies data analysis to the supply chain: measuring inventory and stockouts, forecasting demand with simple, well-evaluated methods, sizing safety stock according to supplier reliability, and choosing a replenishment policy. All calculations use pandas and the standard library, on fictional data.
What you will be able to do
- Compute coverage, turnover and ABC classification of inventory and spot overstock
- Simulate inventory day by day and measure service level, stockouts and OTIF
- Build and evaluate a moving-average forecast without data leakage
- Evaluate suppliers and compute safety stock and reorder point
- Compute an economic order quantity and compare replenishment policies by simulation
Syllabus: 5 lessons
- Compute the value of inventory, its coverage in days and its turnover
- Spot overstock with a coverage threshold
- Classify items into A, B and C according to their weight in consumption
- 02
Stockouts and Service Level
- Simulate inventory evolution day by day and spot stockouts
- Compute quantity service level, stockout days and lost sales
- Measure the OTIF rate (on time and in full) of orders
- 03
Forecasting Demand: Naive, Moving Average and Exponential Smoothing
- Build a naive forecast and a moving-average forecast without data leakage
- Evaluate a forecast with the mean absolute error (MAE) and the bias
- Choose a moving-average window and know the limits of the method
- 04
Supplier Lead Times, Safety Stock and Reorder Point
- Measure the average lead time, variability and reliability of suppliers
- Compute a safety stock for a target service level
- Derive the reorder point and compare suppliers
- 05
Simple Optimization: Economic Order Quantity and Replenishment Policy
- Compute the economic order quantity (EOQ formula) and the associated total cost
- Simulate a reorder-point replenishment policy
- Compare policies on service delivered and average stock
Economic order quantityOrdering and holding costsReplenishment policy (s, Q)
Final project
Inventory improvement plan for a fictional distributor
Using data generated in Python (fixed seed) for about thirty items of a fictional distributor: weekly demand history with a few stockouts, stocks, unit costs and deliveries from three suppliers, build an improvement plan. Classify items by ABC, measure coverage and stockouts, choose a forecasting method for each A item, compute safety stocks, reorder points and economic order quantities, then validate your settings by simulation.
Deliverables
- A reproducible Python script or notebook
- The ABC classification crossed with stock value and the list of overstocks
- The measurement of stockouts, service level and OTIF, with correction for censored demand
- The comparison of forecasting methods (MAE, bias) for A items
- The table of suppliers (lead time, standard deviation, reliability) and of replenishment parameters
- A simulation comparing current and proposed settings, with a note of recommendations
Assessment criteria
- Forecasts only use the past and are evaluated over a common period against a naive benchmark
- Time units are consistent in the safety stock and reorder point calculations
- The simulation accounts for orders in progress and lead times
- Recommendations explicitly trade off service level against average stock
- Assumptions and limits (seasonality, stockouts, discounts) are written down
Frequently asked questions
Do I need to know time series to take this course?
No. The course uses simple methods (naive, moving average, exponential smoothing) and stresses how to evaluate them. The Time Series and Forecasting course (DS6) goes further for those who want it.
Can the safety stock formulas be used in a real company?
Yes, they are classic formulas, provided you respect their assumptions (independent variations close to a normal distribution, consistent units) and validate them by simulation on historical data.
Does the course use management software?
No. Calculations are done in Python with pandas and the standard library, with an equivalent Excel task for the moving average, so you understand what planning software does.
Is the data real?
No, all the data is fictional and built in the code so that every result can be verified.
This course, told for your job
Say what you want to be able to do: “Supply Chain and Forecasting” takes its place in your path, with examples from your field. The concepts stay those of the course.