DataCareerHubLearn · Practice · Progress
Explore learning
DataCareerHub
Explore learning
Learning paths/Data Analyst Foundations
Module 04 · Data exploration

Data quality & exploratory analysis.

Profile imperfect data, document corrections, explore distributions and explain why an overall trend can hide a different story.

7 published lessons340 min suggested studyCSV practice & evidence
Start the first lesson →Get the practice files ↓
The same case, new skills

Northstar Shop

From an imperfect extract
to a defensible explanation.

2 reporting months44 verified orders
Fictional learning case · EUR · No personal data
What you will produce

An evidence pack.
A clearer explanation.

Profile and prepare the extract, review distributions and segment weights, then explain the result with coverage and sensitivity checks. Every lesson includes an explanation, practical task and worked solution.

Before you start: Complete Module 1’s metric and grain concepts. Use a spreadsheet or another local analysis tool that can open UTF-8 CSV files. Modules 2 and 3 offer useful Excel and SQL preparation; Python is optional, not required.

Allow about six hours including exercises. The suggested times are flexible. Keep your work locally; no upload or payment is required.

This module introduces a separate April–May teaching extract: 46 raw records and 44 verified orders. Use any local spreadsheet for the exercises. The optional worked Python source reproduces the checked outputs without third-party packages or a website connection.

A connected learning journey

7 lessons. One complete piece of work.

Explanation + practice + solution
01

Define the extract and build a useful data profile

Define the population, grain, period and units before exploring.

40 min02

Turn data-quality findings into explicit rules

Distinguish validity, completeness, uniqueness and consistency.

45 min03

Prepare data with an auditable correction log

Preserve raw values and document authorised corrections.

50 min04

Explore distributions without deleting inconvenient values

Calculate and interpret mean, median and quartiles.

55 min05

Explain an aggregate trend using segments and weights

Reconcile channel summaries to monthly totals.

55 min06

Keep missingness, coverage and denominators visible

Report observed delivery measures with their eligible counts.

45 min07

Deliver a reproducible exploratory evidence pack

Combine preparation, summaries and graphics into a coherent analysis.

50 min
Everything for the exercise

Imperfect data.
A practical investigation.

Keep the raw extract unchanged. Profile and prepare your own copy, document decisions in the quality log, then compare with the checked outputs after attempting each exercise.

CSV values are in EUR, not cents. Blank delivery fields are missing measurements. The verified high-value order remains in the baseline; scenarios are labelled separately. All records are fictional.

Raw practice extract46 records with duplicate groups and deliberate field issues.
Download ↓
Verified comparison data44 unique fictional orders; delivery blanks remain missing.
Download ↓
Learner worksheetSeven exercises, deliverables and a ten-point self-review.
Download ↓
Setup & data dictionaryField meaning, units, populations and calculation conventions.
Download ↓
Quality-log templateRecord your action and authority for eight investigation entries.
Download ↓
Worked quality logSource-backed actions, including the conflicting duplicate.
Download ↓
Checked monthly resultsCompleted metrics and observed delivery coverage.
Download ↓
Checked channel resultsCounts, values, AOV and order-share weights.
Download ↓
Worked exercise solutionsCalculations, interpretation and a complete evidence note.
Download ↓
Optional reproducible Python analysisReadable standard-library source. Save locally as .py; no dependencies.
Download ↓
Before moving on

Can someone reproduce your answer?

Use the final lesson's 10-point rubric to review your work. A suggested target is 8 points, with no zero in population definitions or preparation provenance and no unsupported causal claim. This is a self-review guide, not an accredited or automatically verified assessment.

Begin Module 04 →

Continue to Module 05: Dashboards & clear explanations →

Self-learning is free. Verify your email for protected practice materials; optional support never affects access.