bachovak/end-to-end-pbi-report-w-ai

This is a portfolio project that demonstrates building a complete Power BI business dashboard using an AI coding assistant, without manually clicking through the Power BI Desktop interface or writing formulas by hand. The dashboard tracks whether suppliers deliver orders on time and in the correct quantity, a metric commonly called OTIF (On-Time-In-Full). The data is fictional, modeled on the kind of analysis a supply chain team at a large manufacturer might do. The project uses Claude Code together with two external tools connected through what the README calls MCP servers. One tool connects the AI directly to Tabular Editor, which is software for building Power BI data models, so the AI could write and test calculation formulas against the actual data before saving them. The second tool handles the report layout, letting the AI add charts, apply a custom color theme, and modify visual settings by editing the underlying report files rather than clicking through menus. A set of domain-specific skill files were also loaded into the session to give the AI context about Power BI conventions and project-specific rules. The workflow started with a written design brief listing what questions each chart should answer. The AI then generated a Python script to create synthetic data, defined the data model structure, wrote roughly twenty calculation formulas covering metrics like rolling three-month averages and month-over-month changes, and built the report page with KPI summary cards, a trend line, a supplier comparison chart, and a callout highlighting an unusual spike in October 2025. The README includes an honest assessment: the data model and formula generation worked reliably, with formulas usually coming out correct on the first or second attempt. The report visuals took more back-and-forth to get right. The author notes the final result is functional but not visually polished, and attributes that to prompt effort rather than a limitation of the tools. This is an educational project. All supplier names and scenarios are fictional and have no connection to the LEGO Group, whose name is used only to set a familiar business context.