About this project

Perspective is a high-performance data visualization and analytics engine capable of handling large-scale streaming datasets. It provides a data-reactive UI as a Custom Element, featuring a virtual-scrolling editable data grid, a WebGL charting engine with over 15 chart types, and tile-based geographic maps. The core query engine is written in C++ and compiled for WebAssembly, Python, and Rust, supporting incremental table updates and reactive joins. It offers a symmetric client/server architecture, allowing the engine to run in-browser, in a Web Worker, or remotely via WebSockets. Key capabilities include: - Support for Apache Arrow, CSV, and JSON formats. - Virtual servers that integrate directly with external engines such as DuckDB, ClickHouse, PostgreSQL, and Polars without requiring ETL. - Integration with Jupyter via a dedicated widget and Python client library. - Bindings for React and server-side support for Node.js, Python (aiohttp, Starlette, Tornado), and Rust.