Perspective — таблицы и графики для потоковых данных

★ 11.3K

Perspective is an open-source data grid, pivot table and charting library for large, real-time and streaming datasets, embeddable in web applications and notebooks. Reach for it when you need a live dashboard that handles millions of rows updating in real time, when you want to give end users drag-and-drop pivot table configuration without custom backend code, when you already have DuckDB, ClickHouse or PostgreSQL and need a UI on top without copying data, or when you're building interactive analytics inside a Jupyter notebook. The core engine is written in C++ and compiled to WebAssembly — including a 64-bit memory64 build for in-browser datasets exceeding 4 GB — with additional runtimes for Python (aiohttp, Starlette, Tornado), Node.js and Rust. Reads and writes Apache Arrow, CSV and JSON; renders 15+ chart types via WebGL plus tile-based geographic maps; provides React bindings and a Jupyter widget via anywidget. Virtual servers translate view configurations into native queries against DuckDB, ClickHouse, PostgreSQL and Polars — no ETL or data copy required. The symmetric client/server API connects to an in-process engine, a Web Worker or a remote WebSocket server. Install via npm (@perspective-dev/client), PyPI (perspective-python) or crates.io. The direction is interactive visualization and analytics, not data pipeline orchestration.