AI Car Theft Analytics

Project intro

When vehicle thefts started climbing in New Zealand in late 2021, data analyst Harpal Sinh Jhala turned 4,553 NZ Police stolen-vehicle reports into a Power BI report that reads like an interactive magazine feature. The dashboard opens with the headline insight — monthly thefts more than doubled, from 464 in October 2021 to 1,053 in March 2022 — then lets you slice by region (Auckland alone accounts for 36%), vehicle type (everyday family cars like stationwagons and saloons are the main targets), and time patterns. Under the hood it's a proper star-schema semantic model with 28 DAX measures across five pages (Overview, Regions, Vehicles, Time Patterns, About), and the report is published to a Microsoft Fabric workspace. The build story is printed right in the dashboard footer: "Designed and built by Harpal Sinh Jhala with Claude Code." It's proof that vibe coding has moved well beyond web apps — Power BI Desktop, Power Query (M), and DAX, an entire BI build orchestrated through Claude Code.