Digital Store Sales BI Analysis — BI Decisions

From Raw Data to Decisions

My objective was to transform fragmented sales data into a unified, interactive BI solution. I led this project end‑to‑end, from data ingestion and cleaning to architecting the model and designing executive‑facing dashboards.

Core Technical Execution

• Data Modeling: Architected an optimized star schema for speed and scalability.
• ETL Pipeline: Built robust Power Query steps to clean/merge multi‑source data (SQL, CSV).
• Advanced Analytics: Authored 30+ DAX measures (Time Intelligence, Campaign ROI, cohorts).
• Governance & Automation: Incremental refresh and RLS with scheduled refreshes.

Key Results & Actionable Insights

• Top Performer: Campaign 6 identified as the leading revenue driver, guiding budget allocation.
• High‑Value Segment: Customers with income $82K–$105K show highest acceptance & sales.
• Agile Decisions: Replaced static reports with self‑service analytics and drill‑through narratives.

• Discovery & KPIs: Align business questions, define success metrics & targets.
• Ingest: Connect SQL/CSV/web sources; set parameters; define dataflows if shared.
• Transform (Power Query): Types, missing values, text cleanup, surrogate keys, currency normalization, calendars.
• Model: Star schema, relationships, direction filters, calculation groups, role tables.
• DAX: Base measures → KPI composites; Time Intelligence (YTD/QTD/YoY), rolling windows, A/B metrics.
• Visual Design: Report themes, consistent cards, bookmarks, what‑if parameters, drill‑through.
• Governance: RLS roles, sensitivity labels, dataset endorsements.
• Deployment: Dev→Test→Prod, scheduled/refresh, incremental policies, alerts.
• Optimization: Model size trimming, query folding, performance analyzer.
• Handover: Readme, data dictionary, runbook for refresh failures.