Mapping applications sometimes need to visualize millions of data points without sacrificing performance or user interactivity. In this talk, we will share a real-world use case where PostgreSQL, PostGIS, and pg_tileserv were combined to create dynamic, filterable aggregation layers. Using function layers, we built a simple stack that supports on-the-fly visualizations, allowing users to explore data with responsive, interactive aggregations. I’ll walk through the design choices, including hex bins, square grids, and dot grids, and discuss the benefits of dynamic filtering for real-time data exploration. Whether you're building a data-heavy GIS application or looking to enhance your digital cartography toolkit, this talk will provide practical insights and code snippets to tackle similar challenges.