Maurice / Mauritius · 6 units · 440 guided minutes · controlled beta
Data and artificial intelligence: concepts, methods, application and evidence
This DASKEREL course develops understanding, visible reasoning and authentic evidence. It is learning support, not a national examination paper or regulated qualification.
Subject specification
Computing, software, networks, data and AI: knowledge, formulae and assessment
Required knowledge strands
- data representation, hardware, operating systems and networks
- algorithms, programming, testing and software design
- data modelling, databases, analytics and AI evaluation
- cybersecurity, privacy, accessibility and responsible systems
Formulae and reasoning frames
- Data rate = data transferred ÷ time
- Storage size = number of items × bits per item
- Availability = uptime ÷ total time × 100%
- Precision = true positives ÷ predicted positives
- Recall = true positives ÷ actual positives
- F1 = 2 × precision × recall ÷ (precision + recall)
- Mean latency = total observed latency ÷ number of observations
Assessment objectives
- explain computing concepts
- apply algorithms and build solutions
- analyse, test and evaluate digital systems
Command words: trace · write · calculate · explain · test · evaluate
Practical and portfolio evidence
- versioned source or design
- test table
- logs
- data card
- security, accessibility and limitation review
Misconceptions explicitly tested: encryption removes the need for access control · a high model accuracy proves fairness and fitness for use · successful execution proves that software is correct
Curriculum and examination-board control
DASKEREL core complete · examination mapping pending
The subject teaching core, formulae, practice and assessment objectives are available. A named examination board or curriculum authority, specification version, jurisdiction, language and assessment format must be verified before board-alignment is claimed.
Country and level beta mapping
Mauritius · Undergraduate and first-cycle higher education
Higher-education alignment is programme-, institution-, regulator- and award-specific; a school examination board must not be reused.
Assessment route: The exact certificate, board, subject code, examination year and assessment format must be recorded before alignment is claimed.
Authority or board candidates
- Mauritius: competent higher-education ministry, regulator or quality-assurance body — applicability to be verified
- Mauritius: awarding university or recognised professional body — programme and institution to be recorded
Academic review required: DASKEREL provides independent beta teaching and assessment support. This record does not claim government, board, university or awarding-body approval, recognition, accreditation or equivalence.
Beta subject-topic matrix (15 subjects)
Academic research and writing
Topics: questions, theory, literature and conceptual frameworks · quantitative, qualitative and mixed methods · sampling, validity, reliability, reflexivity and ethics · analysis, argument, peer review, impact and defence
Formulae/reasoning: Response rate = completed responses ÷ eligible invitations × 100% · Mean = Σx ÷ n; proportion = category count ÷ total · Effect estimate = outcome in comparison groups under a stated design · Claim strength is limited by design, measurement, sample and uncertainty
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Law and governance
Topics: institutions, jurisdiction, authority and sources · rights, duties, liability, remedies and procedure · constitutional, public, criminal, civil and commercial contexts · evidence, interpretation, policy, ethics and access to justice
Formulae/reasoning: Issue → applicable authority → application to material facts → limited conclusion · Policy option score = Σ(weight × evidenced rating), with weights and uncertainty disclosed · Participation rate = participants ÷ eligible population × 100% · Proportionality review = legitimate aim + suitability + necessity + balanced impact
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Finance and accounting
Topics: source documents, double entry, ledgers and trial balance · adjustments, financial statements and cash flow · costing, budgeting, variance and decision support · controls, audit evidence, ethics and reporting quality
Formulae/reasoning: Assets = liabilities + owner equity · Profit = revenue − total expenses · Gross margin = gross profit ÷ revenue × 100% · Current ratio = current assets ÷ current liabilities · Receivables days = trade receivables ÷ credit revenue × period days · Contribution = selling price − variable cost · Break-even units = fixed cost ÷ contribution per unit · Simple interest I = Prt
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Economics and business
Topics: stakeholders, enterprise, objectives and business models · markets, demand, supply, elasticity and economic context · operations, people, quality, finance and risk · strategy, marketing, innovation, sustainability and performance
Formulae/reasoning: Revenue = price × quantity · Profit = revenue − total cost · Market share = business sales ÷ total market sales × 100% · Percentage change = change ÷ original × 100% · Labour productivity = output ÷ labour input · Capacity utilisation = actual output ÷ maximum output × 100% · Price elasticity of demand = % change in quantity demanded ÷ % change in price
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Computing and software
Topics: data representation, hardware, operating systems and networks · algorithms, programming, testing and software design · data modelling, databases, analytics and AI evaluation · cybersecurity, privacy, accessibility and responsible systems
Formulae/reasoning: Data rate = data transferred ÷ time · Storage size = number of items × bits per item · Availability = uptime ÷ total time × 100% · Precision = true positives ÷ predicted positives · Recall = true positives ÷ actual positives · F1 = 2 × precision × recall ÷ (precision + recall) · Mean latency = total observed latency ÷ number of observations
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Data and artificial intelligence
Topics: data representation, hardware, operating systems and networks · algorithms, programming, testing and software design · data modelling, databases, analytics and AI evaluation · cybersecurity, privacy, accessibility and responsible systems
Formulae/reasoning: Data rate = data transferred ÷ time · Storage size = number of items × bits per item · Availability = uptime ÷ total time × 100% · Precision = true positives ÷ predicted positives · Recall = true positives ÷ actual positives · F1 = 2 × precision × recall ÷ (precision + recall) · Mean latency = total observed latency ÷ number of observations
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Engineering and construction
Topics: design requirements, drawings, tolerances and materials · forces, motion, energy and power transmission · electrical, electronic, control and measurement systems · manufacture, construction, maintenance, safety and quality
Formulae/reasoning: Scale factor = drawing length ÷ actual length · Moment M = force F × perpendicular distance d · Stress σ = force F ÷ cross-sectional area A · Work W = Fs and power P = W ÷ t · Ohm’s law V = IR and electrical power P = VI · Efficiency = useful output ÷ total input × 100% · Factor of safety = failure load ÷ allowable working load
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Mathematics and physical sciences
Topics: measurement, motion, forces and momentum · work, energy, power and thermal systems · electricity, circuits, fields and electromagnetism · waves, radiation, matter and space
Formulae/reasoning: Speed v = distance s ÷ time t · Acceleration a = change in velocity Δv ÷ time Δt · Resultant force F = ma · Momentum p = mv · Work W = Fs and power P = E ÷ t · Potential difference V = IR; electrical power P = VI; energy E = Pt · Density ρ = m ÷ V; pressure p = F ÷ A · Wave speed v = fλ · Efficiency = useful output ÷ total input × 100%
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Chemistry and biological sciences
Topics: cells, organisation, transport and bioenergetics · inheritance, variation, evolution and biotechnology · homeostasis, infection, health and populations · ecology, cycles, biodiversity and human impact
Formulae/reasoning: Magnification = image size ÷ actual size · Percentage change = change ÷ original × 100% · Rate = change in quantity ÷ time · Population density = number of organisms ÷ sampled area or volume · Risk rate = events ÷ population at risk over a stated period · Surface-area-to-volume ratio = surface area ÷ volume
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Public health and social care
Topics: cells, organisation, transport and bioenergetics · inheritance, variation, evolution and biotechnology · homeostasis, infection, health and populations · ecology, cycles, biodiversity and human impact
Formulae/reasoning: Magnification = image size ÷ actual size · Percentage change = change ÷ original × 100% · Rate = change in quantity ÷ time · Population density = number of organisms ÷ sampled area or volume · Risk rate = events ÷ population at risk over a stated period · Surface-area-to-volume ratio = surface area ÷ volume
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Education and learning
Topics: questions, theory, literature and conceptual frameworks · quantitative, qualitative and mixed methods · sampling, validity, reliability, reflexivity and ethics · analysis, argument, peer review, impact and defence
Formulae/reasoning: Response rate = completed responses ÷ eligible invitations × 100% · Mean = Σx ÷ n; proportion = category count ÷ total · Effect estimate = outcome in comparison groups under a stated design · Claim strength is limited by design, measurement, sample and uncertainty
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Social sciences and development
Topics: questions, theory, literature and conceptual frameworks · quantitative, qualitative and mixed methods · sampling, validity, reliability, reflexivity and ethics · analysis, argument, peer review, impact and defence
Formulae/reasoning: Response rate = completed responses ÷ eligible invitations × 100% · Mean = Σx ÷ n; proportion = category count ÷ total · Effect estimate = outcome in comparison groups under a stated design · Claim strength is limited by design, measurement, sample and uncertainty
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Languages, media and communication
Topics: reading, inference, comparison and source evaluation · language, structure, genre, context and interpretation · planning, drafting, cohesion, accuracy and editing · speaking, listening, rhetoric, media and audience
Formulae/reasoning: Analytical paragraph = claim + precise evidence + method + explanation + contextual qualification · Source evaluation = provenance + purpose + evidence + corroboration + limitation · Comparative judgement = shared criterion + evidence from each text + similarity or difference + reasoned significance · Argument evaluation = claim + warrant + evidence quality + counterargument + qualified judgement
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Agriculture, environment and sustainability
Topics: place, scale, space, population and development · physical processes, hazards, climate and ecosystems · resources, agriculture, cities, economies and inequality · fieldwork, maps, geospatial evidence and sustainable decisions
Formulae/reasoning: Population density = population ÷ area · Rate of change = change ÷ time · Percentage change = change ÷ original × 100% · Mean = Σx ÷ n and range = maximum − minimum · Irrigation volume = area × water depth · Gross margin = output value − specified variable costs
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Arts, design and architecture
Topics: context, visual or performing language and critical references · research, ideation, experimentation and technique · composition, development, production and presentation · annotation, critique, refinement and personal response
Formulae/reasoning: Scale = representation size ÷ actual size · Proportion and rhythm are evidenced through measured relationships, not treated as universal rules · Evaluation = intention + process evidence + audience response + limitation + revision · Design decision = requirement + explored alternatives + evidence + trade-off + refinement
Comparison: DASKEREL beta drafted; current official specification comparison and academic review required.
Assessment and practical-requirement matrix
DASKEREL beta format: 225 minutes · 15 constructed-response tasks · 150 marks.
Timed, original constructed-response tasks using scenarios, supplied evidence, calculations or reasoned judgements; no answer-option guessing.
Task-specific criteria, transparent workings, evidence quality, limitations and justified conclusions; official grade boundaries are not inferred.
- Academic research and writing: protocol · source and decision log · ethical review · analysis trail · limitations and defence record
- Law and governance: authority table · case or policy brief · fact-to-rule analysis · counterargument · jurisdiction and advice boundary
- Finance and accounting: source register · journal and ledger trail · reconciliation · statement workings · control and reviewer checklist
- Economics and business: stakeholder map · calculation table · option appraisal · sensitivity analysis · conditional recommendation
- Computing and software: versioned source or design · test table · logs · data card · security, accessibility and limitation review
- Data and artificial intelligence: versioned source or design · test table · logs · data card · security, accessibility and limitation review
- Engineering and construction: controlled drawing · calculation sheet · simulation or supervised test · hazard analysis · inspection and revision record
- Mathematics and physical sciences: labelled diagram · raw and processed data · uncertainty record · graph and gradient · risk-controlled method
- Chemistry and biological sciences: method and variables · sampling rationale · raw observations · calculation and graph · ethical and uncertainty statement
- Public health and social care: method and variables · sampling rationale · raw observations · calculation and graph · ethical and uncertainty statement
- Education and learning: protocol · source and decision log · ethical review · analysis trail · limitations and defence record
- Social sciences and development: protocol · source and decision log · ethical review · analysis trail · limitations and defence record
- Languages, media and communication: annotated extract · quotation map · planned and revised response · comparative table · recorded or transcribed presentation
- Agriculture, environment and sustainability: annotated map · sampling plan · data table and graph · stakeholder comparison · sustainability appraisal
- Arts, design and architecture: source-attributed research · sketchbook or process log · experiments · critique response · final presentation and evaluation
Official sources and source dates
No current official specification is recorded for this country and level. Authority identification, official URL, version, effective year and scope remain required before comparison or public alignment claims.
Evidence required for a public alignment claim
- Official authority or awarding-body name, jurisdiction and legal competence
- Current specification, programme or standard URL, publication date, version and effective examination year
- Subject-by-subject and topic-by-topic coverage matrix, including prerequisites and exclusions
- Assessment objectives, command words, paper structure, duration, marks, grading and permitted resources
- Required practicals, coursework, fieldwork, projects, oral work, prescribed texts and safety constraints
- Accessible assessment and reasonable-adjustment rules from the competent authority
- Original DASKEREL mock paper, mark scheme, moderation record and anti-guessing quality review
- Named academically competent reviewer, conflicts declaration, decision, review date and change log
Lesson record and offline copy
mauritius:undergraduate:data-and-artificial-intelligence · Record opened 2026-10-29
- Content origin
- DASKEREL original beta
- Content owner
- DASKEREL
- Reuse rights
- Internal rights review pending
- DASKEREL internal lesson review
- Pending
- Country or level reference records
- 0
- Academic reviewer
- Not recorded; review required
- Language reviewer
- Not recorded; review required
This lesson needs a country-specific terminology and language review before local alignment is claimed.
Use your browser's print dialog to save this page for offline reading. Check the record date when you return online; saved copies do not update automatically.
Six-unit course sequence
Unit 1 · Foundations and precise vocabulary
70 minutes
Build a connected foundation in data and artificial intelligence before attempting assessment tasks.
Study notes
- Study system components, data flows, algorithms, representations, failure modes, security controls and accessibility requirements.
- Build a connected map of these Computing, software, networks, data and AI strands: data representation, hardware, operating systems and networks; algorithms, programming, testing and software design; data modelling, databases, analytics and AI evaluation; cybersecurity, privacy, accessibility and responsible systems.
- Create a glossary that defines each term in context and tests this misconception: encryption removes the need for access control
- Connect the new ideas to prerequisite knowledge, then identify any gap that must be repaired before progression.
Worked application
Deconstruct the lesson’s worked model using the first strand, “data representation, hardware, operating systems and networks”. Label every concept, assumption and decision, then explain why each step belongs to computing, data and digital systems.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “Data rate = data transferred ÷ time” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: encryption removes the need for access control?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Unit 2 · Methods, procedures and transparent reasoning
70 minutes
Learn a repeatable data and artificial intelligence method and know when it is appropriate.
Study notes
- Use this discipline sequence: decompose the problem, define inputs and expected outputs, build in an authorised sandbox, test normal and failure cases, and preserve a rollback path.
- Apply these subject methods: decompose the problem and define acceptance criteria; trace an algorithm using representative and boundary inputs; test expected, invalid, failure and recovery paths; document data provenance, bias, security and human-review boundaries.
- Use and define the required relationships where relevant: Data rate = data transferred ÷ time; Storage size = number of items × bits per item; Availability = uptime ÷ total time × 100%; Precision = true positives ÷ predicted positives; Recall = true positives ÷ actual positives; F1 = 2 × precision × recall ÷ (precision + recall); Mean latency = total observed latency ÷ number of observations.
- Separate information supplied by the task from assumptions introduced by the learner.
- Keep units, citations, definitions, intermediate working and rejected alternatives visible so another person can audit the reasoning.
Worked application
Rework the model with one changed assumption. Select and define an appropriate relationship or reasoning frame from “Data rate = data transferred ÷ time”, predict the consequence before calculating or redrafting, and reconcile the prediction with the result.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “Storage size = number of items × bits per item” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: a high model accuracy proves fairness and fitness for use?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Unit 3 · Guided and independent practice
70 minutes
Move from supported examples to unfamiliar data and artificial intelligence problems.
Study notes
- Complete one imitative task, one task with changed data or context, and one unfamiliar transfer task.
- Practise the command words explicitly: trace, write, calculate, explain, test, evaluate.
- Check performance against these objectives: explain computing concepts; apply algorithms and build solutions; analyse, test and evaluate digital systems.
- After each attempt, classify errors as knowledge, method, evidence, calculation, communication or checking errors.
- Revise the response from feedback and record exactly which change improved the evidence.
Worked application
Compare two plausible responses to the same data and artificial intelligence problem. Use the relevant command word and assessment objective to identify the stronger reasoning, then repair the weaker response without copying the stronger wording.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “Availability = uptime ÷ total time × 100%” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: successful execution proves that software is correct?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Unit 4 · Applied case and decision-making
70 minutes
Apply data and artificial intelligence to a realistic but controlled context.
Study notes
- Use an authorised, privacy-preserving digital service or simulation with explicit users, risks and acceptance criteria.
- Produce appropriate practical or source evidence: versioned source or design; test table; logs; data card; security, accessibility and limitation review.
- Map stakeholders, constraints, benefits, risks and information that remains missing.
- Judge the result using correctness, traceability, security, privacy, accessibility, maintainability and evidence of testing rather than appearance alone.
Worked application
Prepare a decision table with at least two defensible options, weighted or reasoned criteria, a sensitivity check and a conditional recommendation grounded in “versioned source or design”.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “Precision = true positives ÷ predicted positives” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: encryption removes the need for access control?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Unit 5 · Evidence, challenge and quality assurance
70 minutes
Test whether a data and artificial intelligence conclusion survives scrutiny.
Study notes
- Use versioned designs, test cases, logs, datasets, screenshots with text alternatives, configuration records and defect histories.
- Actively test these known misconceptions: encryption removes the need for access control; a high model accuracy proves fairness and fitness for use; successful execution proves that software is correct.
- Distinguish observation, source statement, inference, calculation, value judgement and authorised decision.
- Challenge the conclusion with the strongest alternative explanation and state what new evidence could overturn it.
Worked application
Audit a deliberately flawed response that includes “a high model accuracy proves fairness and fitness for use”, a source problem, a hidden assumption, a method error and an overconfident conclusion; correct each fault and document its effect.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “Recall = true positives ÷ actual positives” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: a high model accuracy proves fairness and fitness for use?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Unit 6 · Synthesis, portfolio and progression
90 minutes
Demonstrate durable data and artificial intelligence mastery through an original artefact.
Study notes
- Plan an artefact for a named audience and assessment purpose, with milestones for research, first attempt, feedback, revision and final verification.
- Show coverage of all four subject strands: data representation, hardware, operating systems and networks; algorithms, programming, testing and software design; data modelling, databases, analytics and AI evaluation; cybersecurity, privacy, accessibility and responsible systems.
- Include a concise explanation for a non-specialist and a technical appendix that preserves the full reasoning.
- Apply this boundary before release: No activity authorises access to another person’s account, production system, confidential data or network without explicit permission.
Worked application
Defend the final recommendation against two questions using “analyse, test and evaluate digital systems”, acknowledge one genuine limitation and identify the next learning step rather than claiming complete certainty.
Guided practice
- Complete the relevant part of the guided practice for data and artificial intelligence; label the command word, selected strand and assessment objective, then annotate every reasoning step.
- Create a second data and artificial intelligence example with different data, audience or constraints, use an appropriate formula or reasoning frame, and solve it independently.
- Use the quality criteria to self-assess, obtain feedback where available, and submit a revised version with a change log.
Mastery checks
- Can you explain the central data and artificial intelligence idea without relying on the worked wording?
- Can you define and apply “F1 = 2 × precision × recall ÷ (precision + recall)” or explain why a different reasoning frame is required?
- Can you detect and correct this misconception: successful execution proves that software is correct?
- Can you identify the evidence, assumptions, limitations and a plausible alternative?
- Can you transfer the method to a new case and justify every important choice?
Portfolio evidence: Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries. For this unit, retain the attempt, feedback, corrections, source record and a short reflection on the next improvement.
Topic-by-topic teaching chapters
Subject chapter
Chapter 1 · data representation, hardware, operating systems and networks
Teaching explanation
- Define the key ideas inside “data representation, hardware, operating systems and networks”, give an example and a non-example, and connect them to the previous chapter rather than learning isolated vocabulary.
- Use this subject method: decompose the problem and define acceptance criteria. Explain why it is appropriate, which assumptions it introduces and which evidence would show that it has failed.
- Apply “Mean data rate = bits transferred ÷ seconds elapsed”. Define every term, preserve units or authority, and interpret the result in the original data and artificial intelligence context.
- Produce versioned source or design, then separate raw evidence, processing, inference and judgement so another learner can audit the conclusion.
- Test and correct this misconception: encryption removes the need for access control
Formula or reasoning frame
Mean data rate = bits transferred ÷ seconds elapsed
Worked solution
- Introductory retrieval only; this case does not establish the depth required for this higher-education module. Supplied data and artificial intelligence case for “data representation, hardware, operating systems and networks”: A fictional log sends 800 bits in 4 seconds. Calculate the mean data rate.
- Select the matching frame, “Mean data rate = bits transferred ÷ seconds elapsed”, and identify the exact quantities or textual details provided by the case before applying it.
- State the condition that makes this frame relevant, including the unit, population, source or decision boundary where applicable; do not supply missing facts as if they were observed.
- Worked solution: Data rate = data transferred ÷ time = 800 bits ÷ 4 s = 200 bit/s. The rate is an average for this transfer, not a guarantee about peak throughput or network security.
- Check the method against “encryption removes the need for access control”; explain which supplied fact supports the conclusion and which further evidence would be needed to generalise it.
Foundation practice
- Reproduce the worked solution for “data representation, hardware, operating systems and networks” with all intermediate steps and units or stated assumptions; then explain one step in your own words.
- Identify which fact in the supplied case is an observation, which is a calculation, and which is a limited interpretation.
Standard practice
- Independent parallel case: A transfer sends 1,200 bits in 6 seconds. Find its mean data rate.
- Explain why “Mean data rate = bits transferred ÷ seconds elapsed” fits this supplied case, and identify one assumption or boundary of the result.
Stretch and transfer
- Compare the worked and independent cases. Change one supplied input or condition, predict the direction of change, and check whether the relationship “Mean data rate = bits transferred ÷ seconds elapsed” still applies.
- Explain which additional versioned source or design would be required before transferring this result to a real data and artificial intelligence decision.
Open chapter answer and marking guidance after attempting the tasks
- Worked-case answer and reasoning: Data rate = data transferred ÷ time = 800 bits ÷ 4 s = 200 bit/s. The rate is an average for this transfer, not a guarantee about peak throughput or network security.
- Independent-case answer and reasoning: 1,200 bits ÷ 6 s = 200 bit/s. State the unit bit/s, not byte/s.
- Foundation practice must preserve the supplied values, show intermediate reasoning and distinguish the stated data from the interpretation.
- For stretch practice, a correct prediction must name the changed input, hold the other assumptions fixed, recalculate or re-evaluate the result, and state why a different condition might invalidate the frame.
Subject chapter
Chapter 2 · algorithms, programming, testing and software design
Teaching explanation
- Define the key ideas inside “algorithms, programming, testing and software design”, give an example and a non-example, and connect them to the previous chapter rather than learning isolated vocabulary.
- Use this subject method: trace an algorithm using representative and boundary inputs. Explain why it is appropriate, which assumptions it introduces and which evidence would show that it has failed.
- Apply “A conditional has a Boolean test, a true branch and a false branch”. Define every term, preserve units or authority, and interpret the result in the original data and artificial intelligence context.
- Produce test table, then separate raw evidence, processing, inference and judgement so another learner can audit the conclusion.
- Test and correct this misconception: a high model accuracy proves fairness and fitness for use
Formula or reasoning frame
A conditional has a Boolean test, a true branch and a false branch
Worked solution
- Introductory retrieval only; this case does not establish the depth required for this higher-education module. Supplied data and artificial intelligence case for “algorithms, programming, testing and software design”: Trace a rule: if score ≥ 50, print 'pass'; otherwise print 'retry'. Use scores 50 and 49.
- Select the matching frame, “A conditional has a Boolean test, a true branch and a false branch”, and identify the exact quantities or textual details provided by the case before applying it.
- State the condition that makes this frame relevant, including the unit, population, source or decision boundary where applicable; do not supply missing facts as if they were observed.
- Worked solution: For 50 the condition is true and the output is 'pass'; for 49 it is false and the output is 'retry'. Testing the boundary 50 is more informative than testing 80 alone.
- Check the method against “a high model accuracy proves fairness and fitness for use”; explain which supplied fact supports the conclusion and which further evidence would be needed to generalise it.
Foundation practice
- Reproduce the worked solution for “algorithms, programming, testing and software design” with all intermediate steps and units or stated assumptions; then explain one step in your own words.
- Identify which fact in the supplied case is an observation, which is a calculation, and which is a limited interpretation.
Standard practice
- Independent parallel case: For the same rule, state the outputs for 0, 75 and the invalid text input 'fifty'.
- Explain why “A conditional has a Boolean test, a true branch and a false branch” fits this supplied case, and identify one assumption or boundary of the result.
Stretch and transfer
- Compare the worked and independent cases. Change one supplied input or condition, predict the direction of change, and check whether the relationship “A conditional has a Boolean test, a true branch and a false branch” still applies.
- Explain which additional test table would be required before transferring this result to a real data and artificial intelligence decision.
Open chapter answer and marking guidance after attempting the tasks
- Worked-case answer and reasoning: For 50 the condition is true and the output is 'pass'; for 49 it is false and the output is 'retry'. Testing the boundary 50 is more informative than testing 80 alone.
- Independent-case answer and reasoning: 0 gives 'retry' and 75 gives 'pass'. 'fifty' needs validation or an error path before the numeric comparison; do not silently count it as 50.
- Foundation practice must preserve the supplied values, show intermediate reasoning and distinguish the stated data from the interpretation.
- For stretch practice, a correct prediction must name the changed input, hold the other assumptions fixed, recalculate or re-evaluate the result, and state why a different condition might invalidate the frame.
Subject chapter
Chapter 3 · data modelling, databases, analytics and AI evaluation
Teaching explanation
- Define the key ideas inside “data modelling, databases, analytics and AI evaluation”, give an example and a non-example, and connect them to the previous chapter rather than learning isolated vocabulary.
- Use this subject method: test expected, invalid, failure and recovery paths. Explain why it is appropriate, which assumptions it introduces and which evidence would show that it has failed.
- Apply “Precision = correct predicted positives ÷ predicted positives; recall = correct predicted positives ÷ actual positives”. Define every term, preserve units or authority, and interpret the result in the original data and artificial intelligence context.
- Produce logs, then separate raw evidence, processing, inference and judgement so another learner can audit the conclusion.
- Test and correct this misconception: successful execution proves that software is correct
Formula or reasoning frame
Precision = correct predicted positives ÷ predicted positives; recall = correct predicted positives ÷ actual positives
Worked solution
- Introductory retrieval only; this case does not establish the depth required for this higher-education module. Supplied data and artificial intelligence case for “data modelling, databases, analytics and AI evaluation”: A fictional classifier predicts 10 positives; 8 are correct. The dataset has 16 actual positives. Find precision and recall.
- Select the matching frame, “Precision = correct predicted positives ÷ predicted positives; recall = correct predicted positives ÷ actual positives”, and identify the exact quantities or textual details provided by the case before applying it.
- State the condition that makes this frame relevant, including the unit, population, source or decision boundary where applicable; do not supply missing facts as if they were observed.
- Worked solution: Precision = 8 ÷ 10 = 0.80; recall = 8 ÷ 16 = 0.50. These values do not establish fairness, safety or suitability for any particular user group.
- Check the method against “successful execution proves that software is correct”; explain which supplied fact supports the conclusion and which further evidence would be needed to generalise it.
Foundation practice
- Reproduce the worked solution for “data modelling, databases, analytics and AI evaluation” with all intermediate steps and units or stated assumptions; then explain one step in your own words.
- Identify which fact in the supplied case is an observation, which is a calculation, and which is a limited interpretation.
Standard practice
- Independent parallel case: A fictional classifier predicts 20 positives; 15 are correct, and there are 25 actual positives. Find precision and recall.
- Explain why “Precision = correct predicted positives ÷ predicted positives; recall = correct predicted positives ÷ actual positives” fits this supplied case, and identify one assumption or boundary of the result.
Stretch and transfer
- Compare the worked and independent cases. Change one supplied input or condition, predict the direction of change, and check whether the relationship “Precision = correct predicted positives ÷ predicted positives; recall = correct predicted positives ÷ actual positives” still applies.
- Explain which additional logs would be required before transferring this result to a real data and artificial intelligence decision.
Open chapter answer and marking guidance after attempting the tasks
- Worked-case answer and reasoning: Precision = 8 ÷ 10 = 0.80; recall = 8 ÷ 16 = 0.50. These values do not establish fairness, safety or suitability for any particular user group.
- Independent-case answer and reasoning: Precision = 15 ÷ 20 = 0.75; recall = 15 ÷ 25 = 0.60. Check the denominators before interpreting performance.
- Foundation practice must preserve the supplied values, show intermediate reasoning and distinguish the stated data from the interpretation.
- For stretch practice, a correct prediction must name the changed input, hold the other assumptions fixed, recalculate or re-evaluate the result, and state why a different condition might invalidate the frame.
Subject chapter
Chapter 4 · cybersecurity, privacy, accessibility and responsible systems
Teaching explanation
- Define the key ideas inside “cybersecurity, privacy, accessibility and responsible systems”, give an example and a non-example, and connect them to the previous chapter rather than learning isolated vocabulary.
- Use this subject method: document data provenance, bias, security and human-review boundaries. Explain why it is appropriate, which assumptions it introduces and which evidence would show that it has failed.
- Apply “Availability = available time ÷ scheduled time × 100%”. Define every term, preserve units or authority, and interpret the result in the original data and artificial intelligence context.
- Produce data card, then separate raw evidence, processing, inference and judgement so another learner can audit the conclusion.
- Test and correct this misconception: encryption removes the need for access control
Formula or reasoning frame
Availability = available time ÷ scheduled time × 100%
Worked solution
- Introductory retrieval only; this case does not establish the depth required for this higher-education module. Supplied data and artificial intelligence case for “cybersecurity, privacy, accessibility and responsible systems”: A fictional service is available for 95 of 100 scheduled hours. Calculate availability.
- Select the matching frame, “Availability = available time ÷ scheduled time × 100%”, and identify the exact quantities or textual details provided by the case before applying it.
- State the condition that makes this frame relevant, including the unit, population, source or decision boundary where applicable; do not supply missing facts as if they were observed.
- Worked solution: Availability = 95 ÷ 100 × 100 = 95%. This measure alone says nothing about confidentiality, accessibility or the severity of the five unavailable hours.
- Check the method against “encryption removes the need for access control”; explain which supplied fact supports the conclusion and which further evidence would be needed to generalise it.
Foundation practice
- Reproduce the worked solution for “cybersecurity, privacy, accessibility and responsible systems” with all intermediate steps and units or stated assumptions; then explain one step in your own words.
- Identify which fact in the supplied case is an observation, which is a calculation, and which is a limited interpretation.
Standard practice
- Independent parallel case: A service is available for 72 of 80 scheduled hours. Calculate availability.
- Explain why “Availability = available time ÷ scheduled time × 100%” fits this supplied case, and identify one assumption or boundary of the result.
Stretch and transfer
- Compare the worked and independent cases. Change one supplied input or condition, predict the direction of change, and check whether the relationship “Availability = available time ÷ scheduled time × 100%” still applies.
- Explain which additional data card would be required before transferring this result to a real data and artificial intelligence decision.
Open chapter answer and marking guidance after attempting the tasks
- Worked-case answer and reasoning: Availability = 95 ÷ 100 × 100 = 95%. This measure alone says nothing about confidentiality, accessibility or the severity of the five unavailable hours.
- Independent-case answer and reasoning: 72 ÷ 80 × 100 = 90%. A separate security and accessibility check is still required.
- Foundation practice must preserve the supplied values, show intermediate reasoning and distinguish the stated data from the interpretation.
- For stretch practice, a correct prediction must name the changed input, hold the other assumptions fixed, recalculate or re-evaluate the result, and state why a different condition might invalidate the frame.
Prerequisites
- 1. Create a diagnostic map of what you already know about data and artificial intelligence, including one secure idea, one uncertain idea and one question that needs evidence.
- 2. Define the starting vocabulary for Computing, software, networks, data and AI and distinguish everyday meanings from discipline-specific meanings.
- 3. Review the prerequisite operations or reasoning needed to use “Data rate = data transferred ÷ time” accurately, including symbols, units, source authority or argument parts as applicable.
- 4. Prepare a notebook or accessible digital document with separate spaces for source records, working, corrections, feedback and final evidence.
Learning objectives
- 1. Explain and connect the four central strands of data and artificial intelligence: data representation, hardware, operating systems and networks; algorithms, programming, testing and software design; data modelling, databases, analytics and AI evaluation; cybersecurity, privacy, accessibility and responsible systems.
- 2. Use these models or reasoning frames accurately and state their limits: Data rate = data transferred ÷ time; Storage size = number of items × bits per item; Availability = uptime ÷ total time × 100%; Precision = true positives ÷ predicted positives; Recall = true positives ÷ actual positives; F1 = 2 × precision × recall ÷ (precision + recall); Mean latency = total observed latency ÷ number of observations.
- 3. Apply and justify the discipline methods: decompose the problem and define acceptance criteria; trace an algorithm using representative and boundary inputs; test expected, invalid, failure and recovery paths; document data provenance, bias, security and human-review boundaries.
- 4. Meet the assessment objectives by responding explicitly to trace, write, calculate, explain, test, evaluate tasks rather than relying on recall alone.
- 5. Produce and quality-assure appropriate evidence: versioned source or design; test table; logs; data card; security, accessibility and limitation review.
Deep explanation
- 1. Computing, software, networks, data and AI is organised around four connected bodies of knowledge. The first, “data representation, hardware, operating systems and networks”, supplies the definitions and foundational relationships needed to interpret a problem precisely. A learner should be able to define the terms, give an example and a non-example, and explain which conditions make the idea applicable.
- 2. The second strand, “algorithms, programming, testing and software design”, develops the mechanisms, structures or patterns that connect evidence to an outcome. Study it by tracing a complete chain from starting information to intermediate reasoning and then to a limited conclusion; do not jump directly from a keyword to an answer.
- 3. The third strand, “data modelling, databases, analytics and AI evaluation”, moves the subject into application. Identify the intended audience, system boundary, stakeholders, constraints, risks and success criteria before selecting a technique. This prevents a technically plausible response from being presented as universally suitable.
- 4. The fourth strand, “cybersecurity, privacy, accessibility and responsible systems”, establishes how claims are tested. Appropriate evidence includes versioned source or design, test table, logs, data card, security, accessibility and limitation review. Keep raw evidence separate from processed results and label observations, sourced statements, calculations, inferences and recommendations.
- 5. Use “Data rate = data transferred ÷ time” as the first quantitative relationship or reasoning frame. Define every symbol or component, state the source of each input, preserve units or citations, show substitution or application, and check the result against the original context. A final number or claim without this chain is not sufficient evidence.
- 6. A second model, “Storage size = number of items × bits per item”, provides an independent way to organise or challenge the response. Compare its assumptions with the first model and explain whether both should agree. Where they do not, investigate data quality, classification, rounding, scope, authority and hidden assumptions before choosing a conclusion.
- 7. A third frame, “Availability = uptime ÷ total time × 100%”, supports transfer to an unfamiliar case. Predict the likely direction or form of the result before completing the method, then reconcile the prediction with the outcome. This makes implausible values, circular arguments and category errors easier to detect.
- 8. The repeatable method is: decompose the problem and define acceptance criteria; trace an algorithm using representative and boundary inputs; test expected, invalid, failure and recovery paths; and document data provenance, bias, security and human-review boundaries. Each stage leaves an audit trail so a teacher, peer or workplace reviewer can reproduce the reasoning and identify exactly where an error or disagreement began.
- 9. Three common traps require deliberate checking: encryption removes the need for access control; a high model accuracy proves fairness and fitness for use; and successful execution proves that software is correct. For each trap, construct a counterexample and name the evidence that would distinguish the misconception from a defensible interpretation.
- 10. Quality in data and artificial intelligence means correctness, traceability, security, privacy, accessibility, maintainability and evidence of testing rather than appearance alone. Apply that standard before release, and respect this boundary: No activity authorises access to another person’s account, production system, confidential data or network without explicit permission. Country terminology, prescribed content, examination format and practical requirements remain subject to current official evidence and qualified academic review.
Worked model
- 1. Step 1 — frame the problem. A learner must make a reasoned data and artificial intelligence decision using supplied evidence. Restate the question, identify the audience and decision boundary, and map it to “data representation, hardware, operating systems and networks” and “data modelling, databases, analytics and AI evaluation”.
- 2. Step 2 — organise inputs. Build a register with the source, date, unit or authority, reliability and intended use of every input. Separate verified evidence from assumptions and mark missing information rather than silently inventing it.
- 3. Step 3 — apply the primary model. Use “Data rate = data transferred ÷ time”; define each term, show every substitution or reasoning link, and retain intermediate working. Check that the result answers the stated question rather than a simpler substitute question.
- 4. Step 4 — verify independently. Apply “Storage size = number of items × bits per item” or test expected, invalid, failure and recovery paths. Compare the outputs, test one boundary case and explain any difference caused by rounding, evidence quality, scope or assumptions.
- 5. Step 5 — challenge the result. Test the response against “encryption removes the need for access control” and the strongest plausible alternative. Change one material assumption and record how the conclusion changes; a robust conclusion should show its sensitivity rather than conceal it.
- 6. Step 6 — conclude conditionally. Judge the response against explain computing concepts; apply algorithms and build solutions; analyse, test and evaluate digital systems. State the supported conclusion, its limitation, the evidence needed to strengthen it and the competent person or authority required before real-world action.
Guided practice
- 1. Build a one-page concept map linking all four strands—data representation, hardware, operating systems and networks; algorithms, programming, testing and software design; data modelling, databases, analytics and AI evaluation; cybersecurity, privacy, accessibility and responsible systems—with definitions, examples and one misconception at each connection.
- 2. Complete a scaffolded example using “Data rate = data transferred ÷ time”. Annotate every symbol, unit, source, assumption and checking step, then write a one-sentence interpretation in context.
- 3. Solve a parallel example using “Storage size = number of items × bits per item” with changed data, audience or constraints. Do not copy the worked wording; justify why the model still applies.
- 4. Compare two plausible responses to a explain task. Use the assessment objectives to identify the stronger evidence chain, then repair the weaker response.
- 5. Design a controlled practical, source-based or simulated activity that produces versioned source or design, test table and logs. Include safety, ethics, accessibility and a method for recording anomalies.
- 6. Audit a deliberately flawed response containing “a high model accuracy proves fairness and fitness for use”, a hidden assumption, an evidence problem and an overconfident conclusion. Correct every fault and explain its effect.
- 7. Create an unfamiliar transfer task using “Availability = uptime ÷ total time × 100%”. Complete it independently, self-assess against explain computing concepts; apply algorithms and build solutions; analyse, test and evaluate digital systems, and revise with a change log.
Understanding and misconception checks
- 1. Explain “data representation, hardware, operating systems and networks” to a non-specialist, then give one case where the explanation would be incomplete without “algorithms, programming, testing and software design”.
- 2. Define every component of “Data rate = data transferred ÷ time” and identify the conditions under which it would give a misleading answer.
- 3. Why is “encryption removes the need for access control” attractive, and what specific evidence would disprove it?
- 4. Distinguish versioned source or design from data card; how should each contribute to a conclusion?
- 5. Respond to the command word “test” using a claim, evidence, explained reasoning, counterargument and qualified judgement.
- 6. Which part of the method—decompose the problem and define acceptance criteria; trace an algorithm using representative and boundary inputs; test expected, invalid, failure and recovery paths; document data provenance, bias, security and human-review boundaries—is most vulnerable to hidden assumptions, and how would you control that risk?
- 7. What must be checked or approved by a teacher, regulator, employer or qualified professional before the learning task becomes real-world action?
Review and progression
- 1. After 24 hours, retrieve the definitions, four strands and “Data rate = data transferred ÷ time” from memory, then correct the response against the lesson rather than merely rereading it.
- 2. After one week, solve a new data and artificial intelligence problem that requires “Availability = uptime ÷ total time × 100%” and one test judgement without using the worked model.
- 3. After one month, revisit the portfolio evidence, identify one claim that now needs stronger support, and add a dated revision explaining what changed and why.
- 4. Retain attempts as well as the final artefact so progress, feedback, authorship and correction are visible to an academic reviewer.
Accessibility and equivalent evidence
- 1. A complete text-first route is available without video, animation or high bandwidth; headings, lists and meaningful link text support navigation.
- 2. Every formula, diagram, map, table or visual representation requires a textual equivalent that preserves the same relationships and assessment demand.
- 3. Audio and oral submissions require a transcript; typed, dictated, enlarged-print, symbol-supported or assistive-technology evidence is acceptable when it demonstrates the same criteria.
- 4. Learners may use a glossary, language scaffold, chunked task sequence and additional processing time without reducing the subject standard being assessed.
Evidence task
Produce a complete data and artificial intelligence teaching portfolio containing a concept map across all four strands, a source and assumption register, two fully worked applications using “Data rate = data transferred ÷ time” and “Storage size = number of items × bits per item”, an independent transfer problem, data card, security, accessibility and limitation review, misconception corrections, feedback, a revision log, an accessible summary and a conditional conclusion that respects professional and country-specific boundaries.
Record sources, assistance received, feedback and revisions. Do not include unnecessary personal or sensitive information.