What you’ll learn
- Design a reproducible computational experiment with documented assumptions.
- Investigate numerical sensitivity and communicate uncertainty without overstating findings.
Proposed course content
5 sections · 11 illustrative lesson records
Research software structure3 lessons
- Introduction to Research Engineering & Scientific Computing
Explore reproducible numerical experiments, computational workflows and careful interpretation of scientific observations.
Public introduction · Video not supplied - Separate experiment configuration, computation and reporting
Separate experiment configuration, computation and reporting. Module: Research software structure.
Enroll to access · Video not supplied - Record environments, data provenance and reproducibility choices
Record environments, data provenance and reproducibility choices. Module: Research software structure.
Enroll to access · Video not supplied
Numerical methods2 lessons
- Explore approximation and numerical stability
Explore approximation and numerical stability. Module: Numerical methods.
Enroll to access · Video not supplied - Review error sources and sensitivity to inputs
Review error sources and sensitivity to inputs. Module: Numerical methods.
Enroll to access · Video not supplied
Simulation and computation2 lessons
- Implement a bounded computational model
Implement a bounded computational model. Module: Simulation and computation.
Enroll to access · Video not supplied - Compare simple and more complex approaches
Compare simple and more complex approaches. Module: Simulation and computation.
Enroll to access · Video not supplied
Experiment design and validation2 lessons
- Define baselines, checks and evaluation criteria
Define baselines, checks and evaluation criteria. Module: Experiment design and validation.
Enroll to access · Video not supplied - Avoid leakage and unsupported causal claims
Avoid leakage and unsupported causal claims. Module: Experiment design and validation.
Enroll to access · Video not supplied
Analysis and technical reporting2 lessons
- Visualize uncertainty and failure cases
Visualize uncertainty and failure cases. Module: Analysis and technical reporting.
Enroll to access · Video not supplied - Prepare a transparent report of assumptions and limitations
Prepare a transparent report of assumptions and limitations. Module: Analysis and technical reporting.
Enroll to access · Video not supplied
Requirements
- Comfortable Python programming and numerical computing experience.
- Working knowledge of linear algebra, statistics and version control.
Tools and resources
- A development computer with resources appropriate to a scoped numerical task.
- Additional compute requirements depend on the proposed experiment and are not confirmed.
- Python and suitable numerical, plotting and notebook tools.
- Version control and an environment-management tool.
Description
Explore reproducible numerical experiments, computational workflows and careful interpretation of scientific observations.
Who this course is for
Experienced programmers planning reproducible computational research.
Example practical project
Proposed example: a reproducible numerical experiment comparing two modeling approaches with sensitivity checks and a technical report.