[News] From weeks of work to days: How I rebuilt two data journalism projects with AI

Award‑winning Italian data journalist Jacopo Ottaviani recounts how he used generative AI to remake two of his seminal data‑visualisation projects in days rather than weeks. The 2012 prison‑death map “Patrie Galere,” originally assembled with BatchGeo, Open Refine and other ad‑hoc tools, was rebuilt in two days with Claude Opus and Sonnet, yielding an open‑source LeafletJS dashboard with bilingual filters, though AI‑generated code required careful cross‑checking to fix clustering and total‑sum errors. A month later, before Anthropic’s Fable 5 model was disabled, Ottaviani used it to create “Strade Mortali,” a Rome road‑safety portal that automatically integrated open data, designed its UI, and invented a weighted “killer‑crossroads index” without prompting. Both cases illustrate AI’s ability to accelerate data‑journalism workflows, produce sophisticated visual assets, and shift reliance from proprietary services to reusable code—while underscoring the need for vigilant validation.