Data Analytics and Data Science

Clean inconsistent text, dates, and categorical values with Excel formulas.

Combine tabular datasets with XLOOKUP while detecting unmatched records.

Summarise business data with PivotTables and calculated fields.

Create dynamic analyses with Excel tables and array formulas.

Reshape and standardise a raw dataset with Power Query.

Implement spreadsheet checks that detect missing, duplicate, and invalid values.

Build a one-variable sensitivity analysis for a business decision.

Design a decision-ready Excel summary with KPIs, trends, and supporting detail.

Build a readable multi-stage analytical query using common table expressions.

Calculate rankings, running totals, and rolling metrics with window functions.

Measure funnel conversion rates from event-level data.

Construct a cohort retention table from customer activity data.

Sessionise timestamped events using SQL window logic.

Create SQL checks for uniqueness, completeness, validity, and referential integrity.

Diagnose a slow analytical query with an execution plan.

Create a reusable analytical dataset as a documented SQL view.

Profile a pandas dataset to detect type, missingness, duplication, and range problems.

Correct data-quality problems with explicit and reproducible pandas transformations.

Merge pandas datasets while validating key cardinality and unmatched records.

Calculate grouped and time-based metrics with vectorised pandas operations.

Reshape data between wide and long formats for analysis and visualization.

Convert repeated notebook logic into a tested Python function.

Organise a notebook so that inputs, transformations, analysis, and outputs are reproducible.

Export a documented analysis-ready dataset for use in Power BI or modeling.

Test whether an exploratory pattern is robust to outliers, small-sample variability, and aggregation choices.

Select a statistical test that matches the research question, variable types, and assumptions.

Diagnose when a statistical test's assumptions are materially violated.

Estimate and interpret a confidence interval and effect size for a business comparison.

Adjust an analysis for the risk created by multiple hypothesis tests.

Estimate the sample size required to detect a specified experimental effect.

Analyse an A/B test using effect size, uncertainty, and decision criteria.

Explain how selection bias, confounding, and measurement error limit a conclusion.

Configure Power Query dependencies, data types, and load settings to feed a Power BI semantic model.

Design a Power BI star schema with appropriate fact and dimension tables.

Configure table relationships and explain their filter-propagation behavior.

Create DAX measures that use row context and filter context correctly.

Implement period-over-period calculations with DAX time intelligence.

Create DAX measures for targets, variance, and variance percentage.

Diagnose incorrect Power BI results caused by model grain or filter context.

Improve a Power BI model by reducing unnecessary columns and cardinality.

Select a chart type that matches a comparison, trend, distribution, relationship, or composition task.

Apply accessible formatting that establishes a clear visual hierarchy.

Configure slicers and visual interactions to support focused exploration.

Create drill-through navigation from an overview to supporting detail.

Design a dashboard page around a defined audience and decision.

Validate dashboard totals and filter behavior against source data.

Present a concise evidence-based narrative that leads from finding to recommendation.

Frame a predictive task by specifying the prediction unit, target, horizon, intended action, and success metric.

Identify target, feature, and temporal leakage in a modeling dataset.

Create a leakage-safe train, validation, and test strategy for a predictive problem.

Build a scikit-learn preprocessing and modeling pipeline.

Choose evaluation metrics that reflect class balance and business error costs.

Compare a simple baseline with candidate models using cross-validation.

Tune model hyperparameters without contaminating final test results.

Select a candidate model using validation evidence while reserving the test set for final evaluation.

Evaluate probability calibration on validation data and apply calibration when it materially improves decision quality.

Select a probability threshold using validation results and explicit business error costs.

Evaluate the resulting decision policy once on held-out test data.

Explain global and individual model behavior with appropriate interpretation methods.

Audit final-model errors across operational and demographic groups to identify weaknesses and material disparities.

Package preprocessing and prediction steps into a reproducible scoring artifact.

Define monitoring checks for input drift, prediction drift, and outcome-based performance.

Write a model card documenting intended use, performance, limitations, and risks.

Create a leakage-safe time-based validation scheme for a forecasting problem.

Decompose a time series into trend, seasonal, and residual components.

Construct and evaluate a seasonal-naive forecast as a baseline.

Build a feature-based forecast using lag and calendar variables.

Compare forecasts with scale-appropriate error metrics.

Create recency, frequency, and monetary-value features from transaction data.

Segment customers with a reproducible clustering workflow.

Evaluate whether customer segments are stable, distinct, and actionable.

Select two feasible public datasets using relevance, quality, scope, and ethical-use criteria.

Write a decision-focused project brief for an analytics case study.

Build a validated SQL analysis layer for the analytics case study.

Produce a focused one-page Power BI dashboard that answers the case study's key questions.

Write recommendations that distinguish evidence, assumptions, limitations, and business implications.

Write a prediction-focused project brief that specifies the target, prediction horizon, intended use, and success metric.

Implement a reproducible baseline-to-candidate modeling pipeline for the data science case study.

Evaluate and interpret the selected model, then write a decision-focused conclusion that states its limitations.

Publish a portfolio project in a recruiter-friendly GitHub repository with a clear commit history and reproducible README.

Convert project results into quantified résumé bullet points without overstating impact.

Deliver a five-minute portfolio walkthrough tailored to a hiring manager.

Solve a timed SQL interview problem while explaining assumptions and edge cases.

Solve a timed pandas data-manipulation problem and validate the result.

Answer an applied statistics interview question using business-oriented reasoning.

Explain model selection, validation, and limitations in a data science interview.

Construct concise STAR stories for collaboration, ambiguity, failure, and stakeholder influence.