How hajarlha/powerbi-sales-dashboard works
Project type and purpose
• This is a self-contained Business Intelligence dashboard built in Power BI Desktop using a CSV dataset (“Global_Superstore2.csv”) to analyze global retail sales (2011–2014).
• It is not a multi-tier web app but an analytical pipeline: data ingestion → transformation → modeling → visualization → user interaction.
Main components
a. Data Source
• The raw CSV file in /dataset/Global_Superstore2.csv.
b. ETL / Data Preparation
• Power Query within Power BI Desktop: data cleaning, type conversions, filter steps.
c. Data Model & Analytical Engine
• In-memory Tabular Model (VertiPaq engine).
• DAX calculation layer (measures: Profit Margin %, Average Order Value).
d. Visualization Layer
• Report canvas: KPI cards, line charts, bar charts, maps, slicers (filters).
e. Output Artifacts
• .pbix file (in /dashboard/orders.pbix).
• Exported PDF (orders.pdf).
• Screenshots folder (for static representation).
f. User Interaction
• Filters/slicers (Date, Region, Country) driving dynamic queries against the model.
g. (Optional) Power BI Service
• If deployed to Power BI Service, include workspace → report → dashboard tiles.
Architectural patterns & design principles
• Layered architecture: Source → ETL → Model → Presentation
• In-memory analytics for fast query performance
• Reusable DAX measures decouple business logic from visuals
• Interactive filters (bi-directional context propagation)
Technologies & frameworks
• Power BI Desktop (Power Query UI, Data Model designer, Report canvas)
• M language (Power Query transformations)
• DAX language (measures and calculated columns)
• VertiPaq storage engine
• PDF export & Power BI Service (optional)
Final checklist
• Components: Data source, ETL, Model, DAX, Visuals, Export/Service
• Detail: Capture every step from CSV import through user-driven filter loops