Be Our Guest
Interested in being a guest on The Data Flowcast? Fill out the form and we will be in touch.
How Airflow and AI Power Investigative Journalism at the Financial Times with Zdravko Hvarlingov
Computational journalism is transforming how news organizations uncover stories from complex data sources. In this episode, we explore how the Financial Times uses Apache Airflow and AI to revolutionize investigative reporting, turning weeks of manual analysis into minutes of automated insights.
Zdravko Hvarlingov, Senior Software Engineer at the Financial Times, joins us to share his insights on building multi-tenant Airflow systems and AI-driven pipelines that surface stories hidden within vast, unstructured datasets.
Key Takeaways:
(00:00) Introduction to computational journalism at the Financial Times.
(02:12) What computational journalism means for day-to-day newsroom work and investigative reporting.
(05:22) Why a shared orchestration platform supports consistent, scalable workflows across teams.
(08:30) Tradeoffs of one centralized Airflow platform versus many separate instances for different use cases.
(11:52) Using data pipelines to structure messy sources for faster journalistic analysis.
(14:14) Turning recurring government disclosures into usable data for ongoing investigations.
(16:03) Applying lightweight machine learning and fuzzy matching to reveal entities and connections.
(18:46) How automation reduces manual effort and shortens time from data to published insight.
(20:41) Practical improvements that make backfilling and pipeline reliability easier for newsroom workflows.
Resources Mentioned:
