Data Analytics

Define observations, variables, units, populations, samples, dimensions, and measures. Covers categorical and numeric data, measurement scales, proxies, missing values, and sources of measurement error.

Trace the field from censuses and statistical quality control through relational databases, spreadsheets, business intelligence, big data, cloud platforms, and AI-assisted analysis. Connect each era to the methods and working practices analysts still use.

Plan surveys, interviews, experiments, transaction logs, sensors, and event tracking so the collected data can answer the intended question. Create a small collection plan with validation rules, consent needs, and a data dictionary.

Work with rows, columns, keys, records, files, and common formats such as CSV, JSON, and Parquet. Reshape messy tables into tidy data without losing meaning or granularity.

Profile completeness, validity, consistency, uniqueness, timeliness, and accuracy. Diagnose missing records, duplicates, impossible values, selection bias, and changes in definitions before trusting a result.

Use formulas, references, lookups, pivot tables, filters, charts, and validation in Excel or Google Sheets. Build an auditable analysis workbook rather than a fragile collection of copied values.

Retrieve, filter, sort, group, and summarize relational data with SQL. Practice handling nulls, dates, text, conditional logic, and common aggregation errors.

Read entity-relationship diagrams and work with primary keys, foreign keys, normalization, and table grain. Design a small relational schema that avoids duplicate facts and ambiguous joins.

Combine and sequence data with joins, subqueries, common table expressions, set operations, and window functions. Solve retention, ranking, running-total, and period-comparison problems in SQL.

Summarize center, spread, shape, position, and association using statistics that fit the data. Interpret distributions, outliers, skew, robust measures, and misleading averages.

Use probability rules, conditional probability, random variables, common distributions, and sampling distributions. Connect sample design and sample size to the uncertainty in an estimate.

Estimate population values with standard errors and confidence intervals, then evaluate claims with hypothesis tests. Interpret p-values, statistical power, effect size, and multiple-testing risk without overstating results.

Model relationships with simple and multiple linear regression, transformations, interactions, and categorical predictors. Check residuals, assumptions, uncertainty, multicollinearity, and the difference between prediction and explanation.

Design randomized experiments and A/B tests with clear units, outcomes, guardrail metrics, sample sizes, and stopping rules. Analyze practical significance while avoiding peeking, novelty effects, and assignment contamination.

Reason about cause and effect using counterfactuals, causal diagrams, confounding, matching, regression adjustment, instrumental variables, difference-in-differences, and regression discontinuity. Judge when observational evidence can support a decision.

Represent uncertain systems with random simulation, bootstrap methods, scenario models, and Monte Carlo analysis. Quantify ranges of possible outcomes instead of reporting a single fragile forecast.

Compare choices using expected value, decision trees, constraints, sensitivity analysis, and optimization. Build simple linear-programming models for allocation, scheduling, and resource-planning decisions.

Correct types, standardize categories, parse dates, handle missing values, remove duplicates, and reconcile inconsistent sources. Build repeatable transformations while preserving raw data and documenting assumptions.

Use summaries, distributions, cross-tabulations, visual checks, and targeted queries to discover structure and anomalies. Keep an analysis log that separates observations, hypotheses, tests, and conclusions.

Choose position, length, color, size, scales, and chart forms that represent data faithfully. Redesign cluttered or misleading graphics with accessible colors, honest axes, and clear comparisons.

Design dashboards around decisions, audiences, refresh timing, filters, drill paths, and metric definitions. Prototype an operational dashboard and test whether users can find and interpret the needed information.

Present findings with a clear claim, relevant evidence, uncertainty, limitations, and recommended action. Adapt written reports, slide decks, annotations, and brief verbal updates to technical and nontechnical audiences.

Use Python variables, collections, control flow, functions, modules, files, and virtual environments for analytical work. Write small scripts that replace repetitive manual steps.

Load, join, reshape, aggregate, and validate datasets with pandas. Build methodical pipelines for dates, categories, missing values, grouped calculations, and export-ready results.

Use R, data frames, the tidyverse, statistical functions, and ggplot2 for reproducible analysis. Translate familiar spreadsheet and SQL tasks into concise R workflows.

Organize projects with notebooks, scripts, configuration files, dependency environments, documentation, and Git. Make an analysis rerunnable by another person without hidden manual steps.

Retrieve data from REST APIs, handle authentication, pagination, rate limits, and nested JSON, and assess when web scraping is lawful and reliable. Store raw responses so later analyses can be reproduced.

Model analytical data with fact tables, dimensions, star schemas, slowly changing dimensions, and OLAP concepts. Choose table grain and aggregation strategies that keep reporting consistent and fast.

Move data through extraction, validation, transformation, loading, scheduling, retries, and dependency management. Compare ETL and ELT patterns and design a pipeline that can recover safely from failures.

Work with elastic storage and compute in platforms such as BigQuery, Snowflake, Redshift, and cloud-native equivalents. Control cost and performance through partitioning, clustering, materialization, workload management, and query inspection.

Manage SQL transformations as tested, documented, version-controlled models using dbt-style workflows. Build staging, intermediate, and mart layers with clear dependencies and deployment practices.

Combine warehouse-style analytics with data-lake storage using lakehouse architectures and open table formats such as Delta Lake, Apache Iceberg, and Apache Hudi. Handle schema evolution, snapshots, transactions, and scalable file layout.

Define governed metrics, dimensions, entities, and calculation rules in a shared semantic layer. Prevent different dashboards from producing incompatible answers to the same business question.

Build governed reports and interactive analysis in tools such as Power BI, Tableau, and Looker. Manage data models, row-level access, refreshes, reusable content, and the tradeoffs of self-service analytics.

Automate checks for schemas, freshness, volume, distribution shifts, broken joins, and metric anomalies. Set service expectations and alerts that help teams distinguish real incidents from harmless variation.

Define producer and consumer expectations for schemas, fields, freshness, compatibility, and change management. Use data contracts to prevent upstream changes from silently breaking downstream analyses.

Catalog datasets, assign ownership, document definitions, classify sensitive fields, and trace lineage from source to report. Apply governance without making useful data needlessly hard to access.

Protect analytical systems with least-privilege access, encryption, secret management, secure sharing, auditing, and safe exports. Recognize injection risks, exposed credentials, unauthorized joins, and other common analyst-side security failures.

Apply data minimization, retention limits, consent, de-identification, aggregation, differential privacy, and controlled clean-room patterns. Evaluate re-identification risk and the privacy tradeoffs of linking datasets.

Identify biased measurements, excluded populations, harmful proxies, unfair comparisons, and incentives created by metrics. Document limitations and design reviews, appeals, and human oversight for high-impact decisions.

Analyze trend, seasonality, cycles, autocorrelation, and structural breaks in time-ordered data. Build and backtest baseline, exponential-smoothing, and ARIMA-style forecasts using time-aware validation.

Measure activation, engagement, conversion, retention, funnels, cohorts, feature use, and experimentation outcomes. Design event schemas and product metrics that connect user behavior to product decisions.

Analyze acquisition channels, campaign performance, customer journeys, segmentation, lifetime value, churn, and attribution. Account for selection effects, delayed conversions, and privacy limits in customer data.

Work with revenue, cost, margin, budgets, cash flow, variance analysis, and financial forecasts. Build decision-ready models with reconciliations, scenario assumptions, and controls over sensitive figures.

Measure capacity, throughput, cycle time, inventory, service levels, quality, and process variation. Apply queueing ideas, control charts, optimization, and forecasting to operational and supply-chain decisions.

Work with coordinates, projections, boundaries, spatial joins, distance, density, and choropleth maps. Analyze location patterns while avoiding projection errors, misleading areas, and exposure of sensitive locations.

Prepare text, extract keywords and entities, measure sentiment, group documents, and create searchable representations. Evaluate language ambiguity, domain drift, multilingual data, and privacy concerns in unstructured text.

Frame classification, regression, clustering, and ranking problems for analytical use. Prepare features, create honest train-validation-test splits, compare classical models, and interpret performance with decision-relevant metrics.

Detect unusual points, shifts, and patterns with statistical rules, control limits, seasonal baselines, and machine-learning methods. Tune alerts around operational cost and investigate whether anomalies are errors, attacks, or real events.

Analyze events as they arrive using streams, windows, event time, late data, and incremental calculations. Design near-real-time metrics and alerts without sacrificing correctness or overwhelming users.

Use coding copilots, natural-language query tools, automated summaries, and generative AI as supervised analytical assistants. Verify generated SQL, calculations, citations, privacy handling, and conclusions rather than treating fluent output as evidence.

Translate stakeholder requests into scope, metric definitions, acceptance criteria, timelines, and explicit assumptions. Handle conflicting definitions, changing requirements, and pressure to support a preferred conclusion.

Deliver scheduled reports, dashboards, alerts, data products, and embedded analytics with ownership and support plans. Monitor freshness, usage, cost, correctness, and decision impact after release.

Complete a realistic project from problem framing and source assessment through cleaning, analysis, validation, communication, delivery, and follow-up. Produce a reproducible repository, documented dataset, tested analysis, decision-ready report, and monitoring plan.

Map roles such as data analyst, business intelligence analyst, product analyst, analytics engineer, and data scientist to their expected skills. Build proof-of-skill projects, prepare for interviews or relevant certifications, find accessible contribution paths, and stay current through documentation, communities, and continued practice.