Free Data Analytics with AI Course — Optional Paid Certificate
Free self-paced data analytics course covering data cleaning, SQL logic, statistics, visualisation, dashboards, forecasting and AI-assisted analysis. Flash cards, checkpoint quizzes and an optional paid completion certificate.
Analysis has value only if it changes an action. Before pulling anything, write the decision it supports, who owns that decision, and what result would push it each way. Requests like 'can you pull the numbers?' should be answered with 'what will you do differently depending on what they say?'.
No decision, no analysis. Define the action first.
Rewrite your next data request as a decision statement with a named owner.
Every expert you admire learned it exactly this way — one scene at a time.
What you will be able to do
- Define a decision-driven question before touching the data
- Clean, join and validate messy real-world data without silently corrupting it
- Think in SQL: filtering, grouping, joins and window functions
- Use descriptive statistics correctly and read distributions, not just averages
- Distinguish correlation from causation and design honest comparisons
- Build charts and dashboards people act on instead of admire
- Apply forecasting and segmentation with realistic expectations
- Use AI to accelerate analysis while keeping the numbers verifiable
Course structure
Foundation — Questions, data quality and the language of data
Where analysis really begins: the decision, the data model, clean joins and the vocabulary every analyst is expected to own.
Advanced — Statistics, experiments and visual truth
Reading distributions, testing honestly, avoiding the classic fallacies, and building charts and dashboards that drive action.
Mastery — Forecasting, AI-assisted analysis and analyst judgement
Prediction with honest error bars, AI as a co-pilot you can audit, governance, and the portfolio that gets you hired.
Full syllabus — every topic covered
30 topic cards across three levels. Each card teaches one idea, ends with a takeaway, and is followed by a checkpoint quiz every five cards.
Level 1: Foundation — Questions, data quality and the language of data
Where analysis really begins: the decision, the data model, clean joins and the vocabulary every analyst is expected to own.
- 1. Start from the decision, not the data — No decision, no analysis. Define the action first.
- 2. Rows, columns and tidy data — One row per observation, one column per variable. Fix shape before analysis.
- 3. Data types and the classic traps — Declare types on import; never let a tool infer identifiers or dates.
- 4. Missing data means something — Measure missingness first; deletion is a decision with consequences.
- 5. Duplicates, keys and grain — Know the grain and the key, and re-check row counts after every join.
- 6. Joins in plain English — Rows disappearing means join type; rows multiplying means a non-unique key.
- 7. How to read SQL even if you don't write it — FROM → WHERE → GROUP BY → HAVING → SELECT → ORDER BY is the mental order.
- 8. Aggregations and the average trap — Report median and percentiles alongside averages, always.
- 9. Metric definitions must be written down — An undefined metric is a future argument. Write the formula down.
- 10. Sanity checks before anybody sees it — Reconcile to a known total and scan for impossible values before you publish.
Level 2: Advanced — Statistics, experiments and visual truth
Reading distributions, testing honestly, avoiding the classic fallacies, and building charts and dashboards that drive action.
- 1. Distributions before conclusions — Plot the distribution before you quote any summary number.
- 2. Variation, and knowing when nothing happened — Compare a change to normal variation before explaining it.
- 3. Correlation is not causation — Ask 'what else could cause both?' before writing a causal sentence.
- 4. A/B testing done honestly — Fix the metric and duration up front; stopping early on a good look is cheating.
- 5. What a p-value and confidence interval mean — Read the interval, not just significance. Wide interval means insufficient data.
- 6. Segmentation and Simpson's paradox — Check the headline inside each major segment before believing it.
- 7. Cohorts and retention — Compare like-aged cohorts, not calendar snapshots.
- 8. Charts that tell the truth — Chart type follows the question; truncated axes and dual axes mislead.
- 9. Dashboards people actually use — One audience, few decision metrics, always a comparison. Then prune.
- 10. Communicating findings to executives — Answer first, evidence second, caveats third — with a quantified action.
Level 3: Mastery — Forecasting, AI-assisted analysis and analyst judgement
Prediction with honest error bars, AI as a co-pilot you can audit, governance, and the portfolio that gets you hired.
- 1. Forecasting basics and honest uncertainty — Beat the naive baseline and always ship a range with the forecast.
- 2. Overfitting and the holdout discipline — Judge models only on data they've never seen, split by time.
- 3. Segmentation and simple predictive scoring — Simple, explainable scoring first; complexity only when it demonstrably pays.
- 4. Using AI as an analysis co-pilot — Let AI write the code; let your database compute the numbers.
- 5. Auditing AI-generated analysis — If you can't see the query and reconcile the total, don't ship the number.
- 6. Privacy, PII and safe data handling — Analyse with the least identifiable data that answers the question.
- 7. Reproducibility and version control — Save the query, script the steps, record the snapshot date.
- 8. Data governance and quality monitoring — Monitor freshness, volume and null rates automatically, with a named owner.
- 9. Stakeholder management and scope — Write the brief, show a rough draft early, refuse decision-less requests.
- 10. The portfolio that gets you hired — Three documented projects with recommendations beat any list of tool logos.
Frequently asked questions
- Do I need maths or coding background?
- No. The course builds from spreadsheet thinking to SQL logic and applied statistics in plain language. You'll understand and be able to explain every concept without heavy maths.
- Which tools does it teach?
- Concepts are tool-agnostic and shown through spreadsheets, SQL and modern BI dashboards, plus how to use AI assistants for analysis safely. Principles transfer to Excel, Google Sheets, Power BI, Tableau, Looker and Python.
- How long does it take?
- About 9 hours across three levels — Foundation, Advanced and Mastery — at your own pace.
- Is this enough to get a data analyst job?
- This course introduces analytics concepts and practice. Build projects that show your reasoning and check the requirements of the roles you want; completing the course does not establish job readiness or guarantee employment.
- What does the certificate include?
- An optional paid completion certificate records your name, course, duration, score, issue date and a verification ID. Review the certificate fee at checkout. It is a record of course completion, not a professional licence or employment guarantee.