Perspective: Interactive Data Visualization for the Browser and Python

Perspective is an interactive analytics and data visualization component designed for large-scale and streaming datasets. It enables developers to build user-configurable reports, dashboards, notebooks, and applications. The library can be deployed as a standalone component in the web browser or integrated into Python and JupyterLab environments.

High-Performance Query Engine The underlying query engine is optimized for speed and memory efficiency. Written in C++, it compiles to WebAssembly and offers native Python support. It provides comprehensive capabilities to read, write, and stream Apache Arrow data. Additionally, its columnar expression language, powered by ExprTK, ensures high-speed data processing and transformation.

Framework-Agnostic UI Perspective’s user interface is delivered as a standard custom element. It can run locally in the browser using WebAssembly or operate through a virtual connection to a WebSocket server running in Python or Node.js.

JupyterLab Integration Perspective includes a dedicated JupyterLab widget and a Python client library, allowing for interactive data analysis directly within notebooks. These workflows can be scaled into production-ready applications using deployment tools like Voila.

Example Projects Example projects demonstrate its versatility across various domains, including fractal market analysis, ray tracing, eviction event tracking, NYPD data exploration, and magical phenomena. Other implementations cover streaming data, COVID-19 statistics, webcam feeds, movie databases, supermarket inventories, Citi Bike usage, and Olympic performance data.