Data Analytics & Computational Intelligence

Frame problems as analytical questions and distinguish evidence from assumption.

• Analytical questions — Analytical questions
• Problem versus analytical question — Problem versus analytical question
• Decision context — Decision context

Understand data structures, variables, measurement and analytical units.

• Variables and observations — Variables and observations
• Measurement scales — Measurement scales
• Data structures — Data structures

Connect questions and hypotheses to appropriate forms of evidence.

• Hypotheses and assumptions — Hypotheses and assumptions
• Evidence types — Evidence types
• Inference boundaries — Inference boundaries

Define scope, stakeholders, constraints and success criteria for an analytical problem.

• Stakeholders and scope — Stakeholders and scope
• Constraints and risks — Constraints and risks
• Success criteria — Success criteria

Identify data sources and assess completeness, consistency and fitness for purpose.

• Structured and unstructured sources — Structured and unstructured sources
• Source fitness — Source fitness
• Data provenance — Data provenance

Transform, clean and structure data for analysis.

• Missing values and anomalies — Missing values and anomalies
• Transformation rules — Transformation rules
• Validation checks — Validation checks

Use descriptive and exploratory techniques to discover patterns and anomalies.

• Descriptive statistics — Descriptive statistics
• Distributions and relationships — Distributions and relationships
• Pattern and anomaly discovery — Pattern and anomaly discovery

Document analytical inputs, provenance, assumptions and evidence quality.

• Lineage and metadata — Lineage and metadata
• Assumptions register — Assumptions register
• Evidence quality record — Evidence quality record

Use statistical reasoning to interpret variation, relationships and uncertainty.

• Variation and distributions — Variation and distributions
• Sampling and estimation — Sampling and estimation
• Statistical relationships — Statistical relationships

Build and evaluate models for prediction and estimation.

• Regression concepts — Regression concepts
• Classification concepts — Classification concepts
• Prediction workflows — Prediction workflows

Assess model performance, robustness and generalisation.

• Train-test evaluation — Train-test evaluation
• Error and performance measures — Error and performance measures
• Robustness and generalisation — Robustness and generalisation

Translate analytical findings into clear decision-relevant communication.

• Analytical narrative — Analytical narrative
• Visual evidence — Visual evidence
• Decision-oriented recommendations — Decision-oriented recommendations

Understand computational approaches for complex analytical problems.

• Computational intelligence concepts — Computational intelligence concepts
• Search and optimisation — Search and optimisation
• Pattern-based computation — Pattern-based computation

Create useful representations and analytical features from data.

• Feature construction — Feature construction
• Encoding and transformation — Encoding and transformation
• Representation quality — Representation quality

Connect data preparation, modelling and execution into repeatable pipelines.

• Data-to-model pipelines — Data-to-model pipelines
• Repeatable execution — Repeatable execution
• Pipeline validation — Pipeline validation

Integrate analytical outputs with workflows, users and decision contexts.

• Decision interfaces — Decision interfaces
• Human-in-the-loop use — Human-in-the-loop use
• Operational feedback — Operational feedback

Identify sources and consequences of bias in data and analytical systems.

• Sources of analytical bias — Sources of analytical bias
• Fairness considerations — Fairness considerations
• Inclusive analytical practice — Inclusive analytical practice

Apply responsible practices for sensitive data and analytical assets.

• Privacy principles — Privacy principles
• Access and protection — Access and protection
• Data stewardship — Data stewardship

Communicate model reasoning, limitations and uncertainty.

• Interpretability — Interpretability
• Uncertainty communication — Uncertainty communication
• Model limitations — Model limitations

Maintain reproducible analytical work and accountable decision records.

• Reproducible workflows — Reproducible workflows
• Audit trail — Audit trail
• Accountability records — Accountability records

Define the applied problem, analytical objectives, evidence plan and success criteria.

• Problem statement — Problem statement
• Analytical objectives — Analytical objectives
• Evidence plan — Evidence plan

Produce the prepared data, analytical artefacts and model outputs.

• Prepared analytical dataset — Prepared analytical dataset
• Model or analytical artefact — Model or analytical artefact
• Supporting documentation — Supporting documentation

Demonstrate evaluation, interpretation and decision relevance.

• Evaluation results — Evaluation results
• Interpretation — Interpretation
• Decision evidence — Decision evidence

Present, defend and reflect on the demonstrated analytical capability.

• Solution walkthrough — Solution walkthrough
• Defence of choices — Defence of choices
• Reflection and next steps — Reflection and next steps