Power BI Pipeline Automation Suite

Building Power BI reports required manually writing DAX measures, configuring visuals, setting up relationships, and structuring the semantic model — a process taking hours per report, requiring specialist Power BI knowledge, and producing inconsistent outputs depending on who built it. Every new dashboard started from scratch with no standardisation, no reuse of logic, and heavy dependency on individuals who understood the data model deeply enough to build it correctly.

Built a fully automated Power BI report generation pipeline that takes a raw semantic model and produces a complete, publish-ready `.pbip` report end-to-end. The system is structured with a Tkinter GUI, covering model scanning, date table generation, calculated columns, analysis table creation, relationship building, DAX measure generation, KPI variant compilation (MoM, YoY, Sparklines), visual layout design, and a final write directly into the Power BI file format. (video pending - in progress) Three orchestration modes were built: a full interactive GUI for step-by-step control, a headless runner for full automation with no user input required, and an AI-driven mode where Google Gemini architects the entire pipeline from a plain-English dashboard intent — replacing steps 2 through 7 with a single intelligent API call. A fourth mode allows users to upload a Word SOP document and have Gemini auto-generate the corresponding measures and visuals directly from it.