Data, machine learning and deep learning
Demand forecasting and time series course
Anticipate demand, sales or consumption with models you can explain.
Forecasting demand well has a direct impact on the bottom line: fewer stock-outs, less tied-up inventory and better planning of purchasing, staffing and cash. It is one of the AI applications whose return is easiest to measure.
This course covers every available tool, from the simplest models (which are often hard to beat) to statistical models, machine learning and the new time series foundation models. We insist on what separates a good forecast from a bad one: temporal validation, the right metrics and communicating uncertainty to whoever makes the decision.
What you will learn
- Analyse time series and their components
- Build baseline and statistical models
- Apply machine learning and time series foundation models
- Evaluate forecasts properly
- Feed forecasts into business decisions
Syllabus
7 modules, from the fundamentals to the final project.
-
Forecasting for the business
- Demand, sales, inventory, cash and consumption
- Horizon, granularity and level of aggregation
-
Analysing series
- Trend, seasonality and cycles
- Holidays, promotions and events
- Missing data and outliers
-
Baseline and statistical models
- Naive and seasonal models
- Exponential smoothing and ARIMA
- Prophet
-
Machine learning for time series
- Lag and calendar features
- Gradient boosting for many series
-
Foundation models
- Pre-trained time series models (Chronos, TimesFM)
- When they work without training
-
Evaluation
- Temporal validation (backtesting)
- MAE, MAPE, bias and value-weighted metrics
- Prediction intervals
-
From model to decision
- Hierarchical forecasting and reconciliation
- Safety stock and scenarios
- Communicating uncertainty
Tools you will practise with
- Python
- pandas
- statsforecast
- Prophet
- LightGBM
- Chronos or TimesFM
Hands-on project
You build a forecasting system on real or sample historical data, comparing statistical, machine learning and foundation models with temporal validation, and present the forecasts with their intervals and their business impact.
Requirements
- Python and pandas (equivalent to the Python for AI course).
Other courses
-
AI for data analysis course (no data science needed)
Draw conclusions from your data with AI, even if you are not a data scientist.
- Duration: 8 hours
- Level: Beginner
-
Computer vision with AI course
Teaching machines to see: inspection, counting, reading documents and more.
- Duration: 20 hours
- Level: Advanced
-
Deep learning with PyTorch course
Understand and build modern neural networks, from tensors to transformers.
- Duration: 30 hours
- Level: Advanced
Where it is applied
-
AI sales and stock forecasting
We use your sales history to anticipate demand for each product and warn you before you run out of stock or overbuy.
Want this course for your team?
We adapt the content, length and format to your company.