Personalized Regional Push Notifications

Summary

This work presents a system for personalized regional push notifications developed at De Persgroep, addressing the challenge of delivering relevant local news content to users at scale. The approach implements machine learning algorithms to automatically select and send push notifications for regional news articles immediately upon publication, learning from both implicit and explicit user feedback to improve personalization over time. The system represents a significant practical contribution to news distribution, scaling from manually curated notifications (~10 per day) to automated personalized delivery of all published content (~500 articles daily), demonstrating how algorithmic approaches can enhance local news engagement while maintaining relevance through geographic and personal preference targeting.