THE ANALYST · CAREER BUILDING SERIES
30-Day Analytics Challenge
Become more than a dashboard builder. Practice framing business problems, designing metrics, writing SQL, building Power BI models, investigating performance, and presenting recommendations like a trusted advisor.
How to use this challenge
- Attempt each task before watching tutorials or looking up solutions.
- Spend 60–90 focused minutes per day. A missed day is a reschedule, not a reset.
- Keep evidence: queries, measures, screenshots, assumptions, and decisions.
- Use synthetic or public data. Never publish confidential employer or customer information.
- At the end of each week, write what you learned and what you would improve.
Your 30 daily missions
Week 1 · Think like a business analyst
Day 01CEO says revenue looks fine but the business may be slowing. Define the problem, list 10 KPIs, 3 analyses, relevant dimensions, 5 possible causes, and a one-slide executive response.
Open Day 1 challenge →Define 10 metrics with business meaning, formula, grain, owner, filters, caveats, and target direction.
Query revenue trends, order counts, AOV, customer cohorts, and returns. Explain the grain and joins.
Compare periods and decompose change by volume, price/value, mix, customer, product, and region. Separate facts from hypotheses.
Find five ways a metric can mislead: denominator shifts, mix changes, averages hiding segments, timing, and data quality.
Sketch a star schema. Define fact grain, dimensions, keys, relationships, and slowly changing attributes where relevant.
Summarize the business question, evidence, caveats, and recommended next action. Include one slide or one-page memo.
Week 2 · Build with SQL and Power BI
Create fact and dimension tables, validate keys, and document relationships.
Build base measures, ratios, distinct counts, and context-aware measures. Avoid unnecessary calculated columns.
Build prior-period, year-to-date, rolling-period, and comparable-period measures. Validate date table behavior.
Use hierarchies and drill-through to move from summary to contributing segments.
Allow users to compare actuals with target and prior period, with clear variance logic.
Design for decisions: concise overview, trends, drivers, exceptions, and action cues.
Remove redundant visuals, test filters, check totals, validate measures, accessibility, and performance.
Week 3 · Investigate and recommend
Identify an unusual KPI movement. Define a baseline, threshold, and false-positive risk.
Decompose the movement, test plausible drivers, and state what evidence would disprove your explanation.
Identify confounders and propose a credible experiment or quasi-experimental approach.
Compare customer, product, region, and channel segments. Explain how outliers affect averages.
Estimate impact in currency or operational units, state assumptions, and show a sensitivity range.
Rank actions by impact, effort, confidence, owner, and measurement plan.
Prepare for five skeptical stakeholder questions. Acknowledge limitations and revise weak claims.
Week 4 · Automate and prove your value
Map the current process, remove manual steps, add validation, and document failure handling.
Profile a dataset, assess missingness and duplicates, summarize drivers, and create a reproducible notebook or script.
Use AI to draft summaries or investigation prompts, but validate every calculation and source. Document privacy guardrails.
Define freshness, completeness, uniqueness, validity, consistency, and reconciliation checks.
In two hours, frame a problem, query data, build a compact analysis, and present a recommendation.
Answer SQL, metric design, stakeholder, prioritization, and analytical judgment questions using specific examples.
Show problem, dataset, method, model, measures, findings, limitations, and actions. Include reproducible evidence.
Turn responsibilities into evidence-backed accomplishment statements. Never invent business outcomes.
Deliver a concise executive presentation and a retrospective: strongest skill, remaining gap, and next 30-day goal.
Definition of done
At the end of 30 days, your portfolio should contain one end-to-end business performance case study, a documented data model, reusable SQL and DAX, a concise executive narrative, and a clear account of limitations and decisions.
Start with Day 01. Do not move to building a dashboard until you can explain the business question and how each KPI helps answer it.