Research Data Analytics Training Course

Participants can identify ethical, privacy and governance controls required for research data.

• Informed consent requirements for primary research data collection
• Purpose limitation and data minimisation applied to research datasets
• Direct and indirect identifiers assessed through disclosure risk
• Pseudonymisation, anonymisation and aggregation compared for analytical use
• Data access roles defined through least-privilege permission principles
• Ethics review records documenting risks, mitigations and approval conditions

Participants can create a structured data management plan for a research project.

• FAIR Principles applied to research data discovery and reuse
• Data management plans covering collection, storage, retention and disposal
• Metadata fields specifying source, owner, refresh date and restrictions
• Data dictionaries defining variables, codes, formats and valid values
• File naming conventions supporting version control and retrieval
• Provenance records linking analytical outputs to source datasets

Participants can select and justify sampling approaches suited to research objectives.

• Target populations, sampling frames and eligibility criteria documented precisely
• Probability sampling: simple random, systematic, stratified and cluster designs
• Sample-size calculations using power, effect size, alpha and precision
• Non-probability sampling biases in purposive, quota and convenience samples
• Weighting adjustments for unequal selection probabilities and non-response
• Recruitment tracking against quotas, response rates and fieldwork risks

Participants can assess whether research instruments produce usable and reliable data.

• Construct validity assessed through indicators, proxies and measurement boundaries
• Likert scales designed with balanced wording and response options
• Questionnaire routing tested for skips, filters and contradictory responses
• Reliability measured using test-retest methods and Cronbach's alpha
• Pilot data reviewed for completion time, ambiguity and missingness
• Collection protocols documenting interviewer instructions and quality checks

Participants can extract defined research datasets using SQL and documented query logic.

• Relational tables interpreted through primary keys and foreign keys
• SELECT, WHERE and ORDER BY used for controlled record extraction
• INNER JOIN and LEFT JOIN applied to linked research tables
• GROUP BY summaries produced for population and response checks
• SQL filters documented against inclusion and exclusion criteria
• Extraction logs recording query versions, dates and row counts

Participants can use Python structures and functions to inspect research data.

• Python notebooks organised into imports, transformations, analysis and outputs
• Variables, lists and dictionaries used to store research parameters
• Tabular datasets loaded into dataframes with explicit data types
• Filtering and sorting records using reproducible selection conditions
• Functions created for repeated checks and standard calculations
• Comments and markdown documenting assumptions, sources and analytical decisions

Participants can assess data fitness before beginning substantive analysis.

• Column profiles measuring completeness, uniqueness, ranges and distributions
• Missing-data patterns classified as MCAR, MAR and MNAR
• Duplicate records detected through keys, matching rules and timestamps
• Outliers investigated using boxplots, z-scores and domain thresholds
• Validity checks comparing values against business rules and codebooks
• Data quality scorecards prioritising remediation by analytical impact

Participants can prepare analysis-ready datasets while preserving an auditable transformation record.

• Missing values handled through deletion, imputation and explicit categories
• Text fields standardised using parsing, casing and controlled mappings
• Dates converted into consistent formats, periods and elapsed durations
• Derived variables calculated from source fields using documented formulas
• Categorical values recoded into analytical groups without losing originals
• Transformation logs recording rules, exceptions, checks and responsible analysts

Participants can identify patterns, anomalies and candidate explanations through exploratory analysis.

• Univariate distributions inspected using histograms, density plots and frequencies
• Cross-tabulations used to compare groups, categories and response patterns
• Scatterplots reveal associations, clusters, nonlinearities and influential observations
• Summary statistics compared across locations, periods and population segments
• Skewness and kurtosis assessed before selecting statistical procedures
• Exploratory findings separated from confirmatory research conclusions

Participants can summarise research data using suitable descriptive and estimation methods.

• Means, medians and modes selected for appropriate measurement scales
• Variance, standard deviation and interquartile range describe dispersion
• Rates, ratios and percentages calculated with defined denominators
• Confidence intervals estimate plausible ranges around sample statistics
• Weighted summaries account for sampling design and response imbalance
• Tables specify populations, time periods, exclusions and missing values

Participants can select, run and interpret common statistical tests appropriately.

• Statistical significance distinguished from practical and policy significance
• t-tests compare mean outcomes between independent or paired groups
• Chi-square tests assess association between categorical variables
• ANOVA compares mean differences across three or more groups
• P-values, confidence intervals and effect sizes interpreted together
• Test assumptions checked before reporting inferential results

Participants can build and interpret regression models for explanatory research questions.

• Linear regression models continuous outcomes using predictor variables
• Logistic regression estimates binary outcome probabilities and odds ratios
• Model specifications reflect theory, hypotheses and confounding factors
• Coefficients interpreted with units, reference categories and uncertainty
• Residual diagnostics identify nonlinearity, heteroscedasticity and influential cases
• Multicollinearity assessed using correlation matrices and variance inflation factors

Participants can identify meaningful groups and patterns within research populations.

• Segmentation criteria defined from research objectives and decision use
• K-means clustering groups observations through distance-based similarity measures
• Cluster counts compared using inertia, separation and interpretability
• Profiles describe segment characteristics, needs, behaviours and outcomes
• Association rules identify co-occurring responses and behavioural patterns
• Segmentation findings tested against sample size and stability limits

Participants can develop a transparent coding approach for qualitative research material.

• Codebooks define themes, inclusion rules, exclusions and examples
• Open coding identifies concepts within interview, focus-group and document data
• Axial coding connects categories, conditions, actions and consequences
• Thematic analysis develops patterned meanings across coded evidence
• Inter-coder agreement assessed through calibration and adjudication discussions
• Analytical memos distinguish participant statements from researcher interpretation

Participants can combine quantitative and qualitative evidence into coherent research findings.

• Convergent designs compare quantitative results with qualitative explanations
• Explanatory sequential designs use interviews to interpret statistical findings
• Joint displays align numerical patterns, themes and integrated conclusions
• Triangulation identifies convergence, complementarity, divergence and unanswered questions
• Mixed-methods claims state evidence sources and conflicting interpretations
• Integration decisions documented within the research analysis plan

Participants can create clear visual evidence displays matched to research questions.

• Chart selection matched to comparison, distribution, relationship and composition questions
• Bar charts use zero baselines and ordered categories appropriately
• Line charts distinguish trends, seasonality and intervention timing
• Scatterplots combine fitted lines, confidence bands and meaningful annotations
• Colour palettes support accessibility, contrast and categorical distinction
• Annotations identify denominators, sources, uncertainty and material limitations

Participants can prepare research analysis for independent review and repeatable execution.

• Reproducible workflows separate raw data, processed data, code and outputs
• Version control practices record changes to scripts and analytical decisions
• Peer review checklists test logic, calculations, assumptions and presentation
• Sensitivity analysis tests findings under alternative assumptions and exclusions
• Validation samples assess whether models generalise beyond development data
• Release checklists confirm confidentiality, accuracy, citations and approvals

Participants can write research reports that connect evidence to bounded recommendations.

• Executive summaries state decision context, findings, implications and actions
• Methods sections document population, sampling, measures and analytical procedures
• Results narratives distinguish observed evidence from interpretation and recommendation
• Limitations address bias, uncertainty, missingness and generalisability explicitly
• Recommendations specify owners, decisions, thresholds and follow-up evidence
• Stakeholder briefings anticipate challenge questions about methods and findings

Participants can build, review and present a complete research analysis pack addressing a defined decision question.

• Research question, hypotheses and decision criteria finalised for capstone analysis
• Source data profiled, cleaned and documented within an audit trail
• Analytical methods selected according to variables, assumptions and research purpose
• Quantitative and qualitative findings integrated into defensible conclusions
• Visual evidence displays prepared with limitations and uncertainty statements
• Peer review applies release checks before stakeholder presentation