forecasting · Mixed async
ARIMA Forecasting Essentials
Fit transparent baseline models, diagnose residuals honestly, and communicate forecast intervals without magical precision.
- Duration
- 6 weeks
- Model focus
- ARIMA
- Cohort
- Evening
- Fee (informational)
- ZAR 1 200
From stationarity intuition to information criteria, this course keeps the workflow modest. You will compare naive benchmarks, document overfitting risks, and produce charts that show uncertainty first.
What is included
- Diagnostics checklist
- Benchmark-first workflow
- Interval communication guide
- Office hours with forecasting coach
- Optional Python track snippets
- Peer review of two forecasts
- Ethics note on presentation of intervals
Outcomes
- Fit and defend a baseline with documented diagnostics
- Contrast ARIMA output against a naive benchmark
- Present intervals without implying false confidence
Kobus Venter
Lead contact
Forecasting coach with a background in applied time series for policy teams.
Questions
Software requirements?
R, Python, or EViews snippets are provided; pick one track and stay consistent.
Math depth?
We use intuition-first explanations; matrix algebra is optional reading.
Limitation?
This is not a machine-learning forecasting course — deep learning modules are intentionally excluded.
Cohort notes
“ARIMA Forecasting Essentials finally made me show the naive benchmark on the same slide — uncomfortable, honest, better meetings.”
Money-Back Policy applies where stated at enrolment.