StarterAfter Excel & Advanced Excel with AI
A tracker that catches its own mistakes
Record income and spending in a workbook that refuses impossible entries: a date in the wrong year, a category that does not exist, a total that no longer adds up.
What you makeA structured workbook, validation rules, a monthly summary and a check that flags anything inconsistent.
The question that tests itPaste in a row that breaks a rule. Show what the workbook does and where the warning appears.
StarterAfter Excel & Advanced Excel with AI
One clean list from three messy ones
Combine three exports of the same people, with different spellings, spacing and duplicates, into a single list you would be willing to send an email from.
What you makeCleaning steps, a duplicate rule you can defend, the merged list and a record of what you removed.
The question that tests itTwo entries might be the same person, or might be two people. Show how you decided, and what it costs to be wrong either way.
StarterAfter Excel & Advanced Excel with AI
Two lists that should agree, and don't
Match a bank export against an invoice list, find the rows that appear in one and not the other, and explain each difference rather than forcing the totals to match.
What you makeA matching approach, an exceptions list, the reconciled total and a note on each unresolved item.
The question that tests itThe totals differ by a small amount. Show whether that is one large error or many small ones.
StarterAfter Excel & Advanced Excel with AI
What the survey actually says
Turn free-text answers into categories you can count, summarise the result and show the quotes behind each category so nobody has to take your word for it.
What you makeA coding scheme, categorised responses, a summary chart and the supporting quotes.
The question that tests itSomeone disagrees with a category. Show your rule and how many answers it moves if you change it.
IntermediateAfter Excel & Advanced Excel with AI
A published dataset that refreshes itself
Connect to a published dataset, shape it into something usable and set it to refresh, so your analysis stops depending on a file someone remembered to download.
What you makeA connection, repeatable preparation steps, a refresh schedule and notes on what to do when the source format changes.
The question that tests itThe publisher adds a column and renames another. Show what breaks and what survives.
IntermediateAfter SQL for Analytics
Find the customers you counted twice
Duplicates quietly inflate every total above them. Find them in SQL, decide which record wins, and measure how much the headline number changes once they are gone.
What you makeDetection queries, a merge rule, a before-and-after comparison and a list of cases needing a human decision.
The question that tests itYour rule merges two genuinely different customers. Show how you would catch that before it reaches a report.
IntermediateAfter SQL for Analytics
Where people drop out
Follow people through the steps of a process and find where most of them stop. The hard part is not the query; it is deciding what each percentage is a percentage of.
What you makeStep-by-step counts, a stated denominator for each rate, a chart and a written finding.
The question that tests itSomeone skips a step and comes back later. Show how your query counts them, and whether that is what you intended.
IntermediateAfter SQL for Analytics
What that discount actually cost
Compare what was sold at a discount with what would probably have sold anyway, and give the business a number with the assumption written next to it.
What you makeA comparison approach, the revenue effect, the assumptions it rests on and a sensitivity check.
The question that tests itChange one assumption. Show how far the answer moves, and say whether it still supports the same decision.
IntermediateAfter SQL for Analytics
Who comes back, and who does not
Group customers by when they arrived and follow each group over time. Retention answers a question a monthly total hides completely.
What you makeCohort query, a retention table, a chart and a written finding with its assumptions.
The question that tests itA group looks worse than the rest. Show whether that is real or an artefact of a short observation window.
IntermediateAfter Excel & Advanced Excel with AI and SQL for Analytics
A monitor that catches bad data first
Checks that run with every refresh: row counts, duplicates, missing values and totals that should reconcile. It tells you before your stakeholder does.
What you makeA check suite, a results log, a failure alert and a note on what to do when a check fails.
The question that tests itA source silently drops ten per cent of its rows. Show which check caught it.
IntermediateAfter Excel & Advanced Excel with AI and Power BI with AI
A monthly pack that builds itself
The reporting routine that took two days becomes a refresh: prepared sources, checked numbers and a pack ready to send.
What you makeRepeatable preparation, validation steps, a scheduled refresh and refresh instructions someone else can follow.
The question that tests itNext month arrives with a renamed column. Show what broke, what warned you and what you changed.
AdvancedAfter SQL for Analytics
Why the report takes four minutes
Take a query everyone complains about, read what the database is actually doing, change one thing and measure. Understanding beats guessing at indexes.
What you makeA baseline measurement, the execution plan, one considered change and the measured result.
The question that tests itYour change helps this query. Show what it costs elsewhere, because an index is never free.
AdvancedAfter SQL for Analytics and Power BI with AI
Did the change actually work?
Two versions ran, one looks better. Decide whether the difference is real, how confident you can be, and what you would tell someone about to spend money on it.
What you makeGroup comparison, the size of the difference, a stated uncertainty and a clear recommendation.
The question that tests itThe result is promising but the sample is small. Show what you would say, and what you would want before deciding.
AdvancedAfter SQL for Analytics and Power BI with AI
Agree what the numbers mean
Three teams report active customers and get three answers. Write the definitions down, show where each existing number came from and get to one agreed set.
What you makeA metric definition for each measure, its source, its known exclusions and a note of what changed once agreed.
The question that tests itA previously published number no longer matches the new definition. Show how you would explain that to the person who used it.
AdvancedAfter Power BI with AI
One dashboard, different audiences
A performance report where each manager sees only their own region, built with row-level security instead of maintaining separate files.
What you makeA security model, roles, tested views for each audience and a handover note.
The question that tests itShow that a regional user cannot reach another region's numbers, including through a filter.
AdvancedAfter SQL for Analytics and Power BI with AI
A forecast you can defend
Project next quarter from history, then show the range around it. A single confident number is usually the least honest answer available.
What you makeA baseline forecast, a stated range, the assumptions behind it and a comparison against what happened.
The question that tests itYour forecast misses. Explain which assumption broke, rather than defending the number.
AdvancedAfter Power BI with AI
A market dashboard that shows its limits
Bring published market or sector data into a model and report the trend clearly, including what the data cannot tell you.
What you makeData model, trend measures, a readable report and a stated list of limitations.
The question that tests itSomeone reads a prediction into your chart. Show what the data supports and what it does not. This is an analysis exercise, not investment advice.