Download PBIX (latest): see Releases.

Production-style Power BI project that turns the Walmart weekly sales dataset into decision-ready insights.
The model is lightweight (fact + dimension), refreshes from CSVs hosted in this repo, and follows PL-300 best practices.
All M (Power Query) and DAX are versioned, and the PBIX is tracked with Git LFS.

• Executive Overview – live KPIs, weekly sales pattern by store type, top store summary
• Profitability Insights – store-type contribution, holiday vs non-holiday profit, weekly profit trend
• Reusable DAX library – curated measures for KPIs, YoY, tooltips, and narrative labels
• Reproducible refresh – Power Query loads raw CSVs directly from GitHub (Anonymous)

• train.csv – weekly sales by Store & Dept
• features.csv – Temperature, Fuel_Price, MarkDown1–5, CPI, Unemployment, IsHoliday
• stores.csv – Store_Type & Size

Source: Kaggle Walmart Sales (files hosted in this repo for reproducible refresh)

• Replace "NA" → null in MarkDown1–5
• Enforce numeric types for MarkDown1–5, CPI, Unemployment, Temperature, Fuel_Price
• Normalize names (e.g., Type → Store_Type)
• Join Sales ↔ Stores, then enrich with Features
• Add helpers: Week_Num, Year, labels/colors for clean visuals
  All query scripts are exported under powerquery/queries/.

Star schema:

• Fact: SalesFact
  (Date, Store, Dept, Week_Num, Weekly_Sales, IsHoliday, Temperature, Fuel_Price, MarkDown1–5, CPI, Unemployment, Store_Type, Size)
• Dim: Stores (Store, Store_Type, Size, formatting fields)
• Relationship: Stores[Store] 1 ─── * SalesFact[Store]

• Total_Sales
• Estimated_Profit (24.5% margin on sales)
• Avg_Weekly_Sales / Avg_Weekly_Profit (context-aware by selections)
• YoY_Sales_Growth (selected Year vs prior Year; blank for first year)
• Top_Store_ID / Top_StoreType + narrative/tooltip helpers

• Measures: dax/measures/core_measures.dax
• Calc columns: dax/calc-columns/fact_and_dim_columns.dax
• Architecture: Docs/Architecture.md
• KPIs: Docs/KPIs.md
• Changelog: Docs/CHANGELOG.md

• Executive Overview – KPIs, weekly pattern, top store type
  View screenshot

• Profitability Insights – profit by store type, holiday vs non-holiday, weekly trend
  View screenshot

  Screenshots reflect the full date range. Key insights below summarize the latest year.

1. Open PowerBI_Files/Walmart_Sales.pbix in Power BI Desktop.
2. If prompted, set Anonymous credentials for https://raw.githubusercontent.com/.
3. Home → Transform data (review) → Close & Apply → Refresh.

• PowerBI_Files/ – PBIX (Git LFS)
• Data/ – train.csv, features.csv, stores.csv (raw inputs)
• powerquery/queries/ – M code for Sales, Features, Stores, Sales_Stores, SalesFact
• dax/measures/ – core_measures.dax
• dax/calc-columns/ – fact_and_dim_columns.dax
• assets/screenshots/ – dashboard images
• Docs/ – architecture/KPIs/changelog

• YoY_Sales_Growth_v2 compares the selected Year to the prior year; first year shows blank by design.

• “Top Store” measures are filter-aware (respect slicers); “Top Store Type” is kept separate on purpose.

  Key insights (latest year in data)

• Latest year: 2012

• Top store: Store 4 (Type A) with $2.00bn in sales.

• Year-over-year: –18% vs prior year.

• Holiday impact: 5% of annual sales (≈ $98.34M).

• Peak trading week: Week 15.

License: MIT — see LICENSE