{"id":109056,"date":"2026-08-04T20:16:16","date_gmt":"2026-08-04T20:16:16","guid":{"rendered":"https:\/\/youzum.net\/reflex-open-sources-xy-a-rust-backed-super-fast-python-charting-library-that-keeps-100-million-point-charts-interactive\/"},"modified":"2026-08-04T20:16:16","modified_gmt":"2026-08-04T20:16:16","slug":"reflex-open-sources-xy-a-rust-backed-super-fast-python-charting-library-that-keeps-100-million-point-charts-interactive","status":"publish","type":"post","link":"https:\/\/youzum.net\/de\/reflex-open-sources-xy-a-rust-backed-super-fast-python-charting-library-that-keeps-100-million-point-charts-interactive\/","title":{"rendered":"Reflex Open Sources XY: A Rust-Backed Super-Fast Python Charting Library That Keeps 100 Million Point Charts Interactive"},"content":{"rendered":"

Reflex<\/a> AI has released XY<\/a>, an Apache-2.0 Python charting library for interactive 2D visualization. Most Python charting stacks create one drawable object per row, so past a few hundred thousand points, render, hover, and zoom degrade. XY moves the work into a native Rust core, sends the browser typed binary buffers instead of JSON, and draws with WebGL2. In the terms of the benchmark, XY holds 0.071 s at 10,000 points and 0.081 s at 100 million<\/strong>. It ships as pip install xy<\/code> and requires Python 3.11 or newer.<\/p>\n

Is it deployable<\/strong><\/h2>\n

XY<\/strong> is early alpha at version 0.0.1<\/strong>. It ships as pip install xy<\/code> and requires Python 3.11 or newer.<\/p>\n

That maps to a clear deployment envelope. Startups and mid-size data teams can adopt it now for internal analytics, notebooks, and shareable artifacts. Regulated enterprises should pilot it rather than put it on a customer-facing critical path. Fit is strongest where row counts are the actual bottleneck: quantitative finance<\/strong> (tick data), genomics and bioinformatics<\/strong> (Manhattan plots, allele-frequency scans), observability and telemetry<\/strong>, astronomy<\/strong>, and geospatial analytics<\/strong>. Here is how to install it in 1 line.<\/p>\n

\n
\n
\n
\n
<\/div>\n
<\/div>\n
<\/div>\n<\/div>\n
Copy Code<\/span>Copied<\/span>Use a different Browser<\/span><\/a><\/div>\n<\/div>\n
pip install xy<\/code><\/pre>\n<\/div>\n<\/div>\n
Explainer<\/strong><\/h2>\n
\n<\/div>\n
\n
How the representation ladder works<\/strong><\/h2>\n<\/p>
XY keeps canonical f64 columns in a ColumnStore<\/code> in Python and picks a rendered representation per trace. Current defaults start M4 decimation above 10,000 rows<\/strong> for long ordered lines, and automatic scatter density above 200,000 points<\/strong>. Density grids default to 512\u00d7384 cells. The docs are explicit that these are pre-1.0 policy thresholds, not API guarantees.<\/p>\n
Because exact values stay in Python, hover, selection, and pick()<\/code> still resolve original rows when the active tier has an exact mapping. Zooming into a narrow window returns exact visible points for a padded aligned window, and nearby pans render from that cached window without another request. Reflex is careful not to overclaim here: density is natively binned and GPU-rendered, not an all-GPU ingest pipeline, and ingest, binning, and decimation still scale with source row count.<\/p>\n
Performance<\/strong><\/h2>\n
The benchmark<\/a> drives every library through a real browser and stops the clock only when the canvas is verified correct and stable across 10 byte-identical frames. Measurements come from one Apple M5 Pro, one run per cell.<\/p>\n
\n\n\n\n\n\n\n\n\n
Points<\/th>\nXY<\/th>\nMatplotlib (WebAgg)<\/th>\nPlotly (scattergl)<\/th>\n<\/tr>\n<\/thead>\n
1M<\/td>\n0.084 s<\/td>\n0.357 s<\/td>\n0.614 s<\/td>\n<\/tr>\n
10M<\/td>\n0.083 s<\/td>\n2.804 s<\/td>\n3.367 s<\/td>\n<\/tr>\n
50M<\/td>\n0.076 s<\/td>\n13.385 s<\/td>\n\u2715<\/td>\n<\/tr>\n
100M<\/td>\n0.081 s<\/td>\n\u2715<\/td>\n\u2715<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n
That is a stated 34\u00d7 speedup at 10M and 177\u00d7 at 50M. Peak Python-side memory at 10M is 0.32 GiB for XY against 0.84 GiB for Matplotlib and 1.86 GiB for Plotly. With density=False<\/code>, XY still draws 100M exact markers in 1.343 s on 5.26 GiB. Reflex also reports rendering the full OpenStreetMap dataset \u2014 10 billion points<\/strong>.<\/p>\n
A 10-million-point interactive scatter exports to 258 KiB<\/strong> of HTML, versus a stated 259 MiB for the Plotly equivalent. The payload stays near 258 KiB from 1M through 100M rows.<\/p>\n
API surface and integration<\/strong><\/h2>\n
Charts are composed declaratively from marks, axes, legends, tooltips, and annotations. Fourteen chart families ship today, including scatter, line, area, histogram, box, violin, ECDF, heatmap, hexbin, and contour. Styling accepts CSS and Tailwind classes through stable DOM slots. For migration, import xy.pyplot as plt<\/code> runs common Matplotlib pyplot code, though the compatibility guide<\/a> notes not everything is supported. A separate reflex-xy<\/code> adapter turns any chart into a Reflex component with no JavaScript or iframe.<\/p>\n
Key Takeaways<\/strong><\/h2>\n