The dataset library
Eleven datasets built for this course — 24,252 rows in total. They are generated rather than scraped, which means they are free of licensing problems and free of personal data, but they are not toy data: they carry real seasonality, genuine margin differences between categories, attrition that correlates with tenure and satisfaction, and in one case a deliberate set of data-quality faults for you to fix.
Every file is plain UTF-8 CSV with a header row, so it loads cleanly into Power BI Desktop, Excel, Python or anything else.
bwb-retail-sales.csv (the course's
main fact table, used from Module 0 to Module 11) and bwb-messy-customers.csv (the
deliberately dirty file that every Power Query lesson uses). The dimension tables
— products, customers, stores, date — join to the sales file to form the star schema
you build in Module 4.Two full years (Jan 2024 – Dec 2025) of order-line transactions for a mid-sized multi-channel retailer selling across six product categories, five regions and three channels. Contains real seasonality — a November peak, a February trough — and category margins that differ enough to make the analysis interesting. This is the course's primary fact table.
| Column | Data type | Meaning |
|---|---|---|
OrderID | Text | Order identifier. One order can have several rows — one per product line. |
OrderDate | Date | Date the order was placed. Use this for time intelligence. |
ShipDate | Date | Date the order shipped. A second date column, for the role-playing dimension lesson. |
CustomerKey | Text | Foreign key to bwb-customers.csv. |
ProductKey | Text | Foreign key to bwb-products.csv. |
StoreKey | Text | Foreign key to bwb-stores.csv. Blank for non-store channels. |
Channel | Text | Online, Retail Store or Partner. |
Quantity | Whole number | Units sold on this line. |
UnitPrice | Decimal | List price per unit before discount. |
Discount | Decimal | Discount rate applied, 0 to 0.30. |
NetSales | Decimal | Revenue after discount. The measure you will sum most often. |
COGS | Decimal | Cost of goods sold for this line. |
Profit | Decimal | NetSales minus COGS. Pre-computed so you can validate your own DAX. |
- Which category generates the most revenue, and is it also the most profitable?
- How did monthly revenue trend across the two years, and where is the seasonal peak?
- Which region has the highest revenue per customer?
- What share of revenue comes from discounted orders, and does discounting improve profit?
- Which products account for the top 20% of revenue?
One year of hospital encounters: department, admission type, length of stay, cost, severity and outcome. Length of stay and cost vary systematically by department, which makes utilisation and cost-per-encounter analysis meaningful rather than noise.
| Column | Data type | Meaning |
|---|---|---|
EncounterID | Text | Primary key — one row per encounter. |
PatientID | Text | Patients can appear more than once. Distinct count matters here. |
AdmitDate | Date | Admission date. |
DischargeDate | Date | Discharge date. |
Department | Text | Seven clinical departments. |
AdmissionType | Text | Emergency, Elective, Referral or Transfer. |
LengthOfStayDays | Whole number | Days between admit and discharge. Zero means same-day. |
PatientAge | Whole number | Age at admission. |
PatientGender | Text | Recorded gender. |
Outcome | Text | Discharge outcome, including 30-day readmission. |
TotalCost | Decimal | Total encounter cost. |
PayerType | Text | Private, public or self-pay. |
SeverityScore | Whole number | 1–5 clinical severity. |
Readmitted30Day | Text | Yes / No flag. |
- Which department has the highest cost per encounter, and is severity the cause?
- What is the 30-day readmission rate, and does it vary by admission type?
- How does length of stay distribute — is the mean misleading here?
- Which departments show the widest cost variation for similar severity?
A year of purchase orders across seven suppliers and four warehouses, with promised versus actual lead times, fill rates and closing stock. Delivery performance genuinely varies by supplier, so supplier scorecards produce a real ranking rather than noise.
| Column | Data type | Meaning |
|---|---|---|
PurchaseOrderID | Text | Primary key. |
OrderDate | Date | Date the PO was raised. |
PromisedDate | Date | Date the supplier committed to. |
ReceivedDate | Date | Date goods actually arrived. |
SupplierKey | Text | Supplier identifier. |
SupplierName | Text | Supplier display name. |
SupplierCountry | Text | Country of origin. |
WarehouseKey | Text | Receiving warehouse identifier. |
WarehouseName | Text | Warehouse display name. |
WarehouseRegion | Text | Region the warehouse serves. |
ProductKey | Text | Joins to bwb-products.csv. |
ProductCategory | Text | Product category. |
QuantityOrdered | Whole number | Units ordered. |
QuantityReceived | Whole number | Units actually received — fill rate is derived from this. |
PurchaseCost | Decimal | Cost of the order. |
PromisedLeadDays | Whole number | Committed lead time. |
ActualLeadDays | Whole number | Actual lead time. |
DelayDays | Whole number | Actual minus promised. Negative means early. |
DeliveryStatus | Text | On Time, Late or Severely Late. |
ClosingStockUnits | Whole number | Stock on hand after receipt. |
SupplierQualityScore | Decimal | Supplier quality rating out of 5. |
- Which suppliers miss their promised lead time most often?
- Does a higher quality score actually correlate with on-time delivery?
- Which warehouse absorbs the most late deliveries?
- What is the fill-rate distribution, and which category is worst affected?
Employee master data for a 1,200-person company: department, level, salary, tenure, satisfaction, performance and attrition. Attrition is not random — it correlates with short tenure, low satisfaction and specific departments, so the analysis produces findings worth acting on.
| Column | Data type | Meaning |
|---|---|---|
EmployeeID | Text | Primary key. |
EmployeeName | Text | Display name. |
Department | Text | Seven departments. |
JobLevel | Text | Associate, Mid, Senior, Lead or Director. |
HireDate | Date | Start date. Tenure is derived from this. |
ExitDate | Date | Blank for current employees — a genuine null to handle. |
Attrition | Text | Yes / No. The target of the analysis. |
AnnualSalary | Whole number | Base salary. |
Education | Text | Highest qualification. |
Gender | Text | Includes a 'Prefer not to say' category — handle it honestly. |
Age | Whole number | Age in years. |
SatisfactionScore | Whole number | 1–5 self-reported survey score. |
PerformanceRating | Whole number | 1–5 manager rating. |
WorkMode | Text | Remote, Hybrid or On-site. |
Region | Text | Office region. |
- Which department has the highest attrition rate, and is the difference material?
- Does satisfaction score predict attrition better than salary does?
- What is the tenure profile of leavers versus stayers?
- Is there a pay gap by level once department is controlled for?
The customer master: region, segment and signup date. Clean, one row per customer, with a proper key. Use it to slice the sales fact by geography and segment.
| Column | Data type | Meaning |
|---|---|---|
CustomerKey | Text | Primary key. Joins to bwb-retail-sales.csv. |
CustomerName | Text | Display name. |
Region | Text | North, South, East, West or Central. |
Segment | Text | Consumer, Small Business or Enterprise. |
SignupDate | Date | First registration date — useful for cohort analysis. |
- Which segment has the highest average order value?
- Do customers who signed up earlier spend more over time?
A complete, gap-free date table covering 2024–2025 with calendar and fiscal attributes. Provided as a reference so you can compare it against the one you build yourself in DAX or Power Query — building your own is the better habit.
| Column | Data type | Meaning |
|---|---|---|
Date | Date | One row per day, no gaps. This is the key column. |
Year | Whole number | Calendar year. |
MonthNumber | Whole number | 1–12. Use this to sort MonthShort correctly. |
MonthShort | Text | Jan, Feb, Mar … Needs Sort By Column set to MonthNumber. |
MonthName | Text | Full month name. |
QuarterNumber | Whole number | 1–4. |
Quarter | Text | Q1–Q4. |
WeekOfYear | Whole number | ISO week number. |
DayName | Text | Monday … Sunday. |
DayOfWeek | Whole number | 1 = Monday. Sort key for DayName. |
IsWeekend | Text | Yes / No flag. |
YearMonth | Text | YYYY-MM, handy for a continuous month axis. |
FiscalYear | Whole number | Fiscal year starting 1 July. |
FiscalMonth | Whole number | 1–12 where 1 = July. |
- How does a fiscal-year view change the picture versus calendar year?
- Which weekday generates the most revenue?
Two years of monthly actual and budget amounts by account and cost centre. Revenue accounts are positive, cost accounts are negative — the sign convention real ledgers use, and the one that breaks naive SUM measures.
| Column | Data type | Meaning |
|---|---|---|
PeriodStart | Date | First day of the month the amounts belong to. |
AccountCode | Text | General-ledger account code. |
AccountName | Text | Human-readable account name. |
AccountType | Text | Revenue, COGS or OpEx — drives the sign convention. |
CostCentre | Text | Owning cost centre. |
BudgetAmount | Decimal | Planned amount for the period. |
ActualAmount | Decimal | Actual amount for the period. |
Variance | Decimal | Actual minus Budget. Pre-computed so you can check your DAX. |
- Which cost centres are consistently over budget?
- What is the gross margin trend by month?
- Does favourable revenue variance offset unfavourable cost variance?
- Which single account drives the largest absolute variance?
A deliberately dirty extract — the kind of file that actually lands in your inbox. It combines two exports, so it contains duplicate rows and duplicate customer IDs. Country is spelled 25 different ways for 6 real countries. Dates use five different formats. Revenue mixes currency symbols, thousands separators and blanks. Status has 7 spellings for 3 real states. Emails are missing, malformed or padded with whitespace. Every Power Query lesson uses this file.
| Column | Data type | Meaning |
|---|---|---|
CustomerID | Text | Intended to be unique — it is not. Duplicates are your first clue. |
FullName | Text | Some entries have double spaces, trailing spaces, or ALL CAPS. |
Email | Text | Blank in some rows, malformed in others, whitespace-padded in others. |
Country | Text | 25 distinct spellings for 6 real countries. |
SignupDate | Text | Loads as text because five date formats are mixed together. |
LifetimeRevenue | Text | Loads as text: currency symbols and thousands separators present. |
Phone | Text | Four different formatting conventions. |
Status | Text | 7 spellings for 3 real states (Active, Inactive, Churned). |
- How many genuinely distinct customers are in this file?
- What is total lifetime revenue once the column is numeric?
- How many customers are Active once status is standardised?
- Which rows should be rejected rather than repaired, and why?
The product catalogue: category, subcategory, list price and unit cost. A clean dimension table with a genuine primary key — exactly what a star schema needs.
| Column | Data type | Meaning |
|---|---|---|
ProductKey | Text | Primary key. Joins to bwb-retail-sales.csv. |
ProductName | Text | Display name for the product. |
Category | Text | One of six top-level categories. |
Subcategory | Text | Finer grouping inside the category. Builds a natural hierarchy. |
UnitPrice | Decimal | Current list price. |
UnitCost | Decimal | Current unit cost. |
- Which subcategories carry the widest price range?
- What is the theoretical margin per category before discounting?
Sales by category and region with twelve month columns spread across the row — the shape a finance team hands you from Excel. Power BI cannot use this directly. Unpivoting it into long format is the whole point of the exercise.
| Column | Data type | Meaning |
|---|---|---|
Category | Text | Product category. |
Region | Text | Sales region. |
Jan-2025 … Dec-2025 | Decimal | Twelve month columns. These need to become two columns: Month and Sales. |
- What does this table look like after Unpivot Other Columns?
- Why can you not build a month trend line from the file as delivered?
Eight physical stores with format and opening year. A small dimension, useful for demonstrating that a dimension does not need to be large to be valuable.
| Column | Data type | Meaning |
|---|---|---|
StoreKey | Text | Primary key. Joins to bwb-retail-sales.csv. |
StoreName | Text | Display name. |
Region | Text | Region the store sits in. |
StoreFormat | Text | Flagship, Standard or Compact. |
OpenedYear | Whole number | Year the store opened — lets you compare mature vs new stores. |
- Do flagship stores out-earn standard stores per square metre of format?
- How do stores opened in 2023–24 compare against mature stores?