New post: Tagging my blog with BERTopic and LLMs. I suspect as token costs increase, we'll see more blended traditional ML/LLM systems. And, in general, they work really well! vickiboykis.com/2026/05/18/t...

Explore yourself at: dtour.dev

High-dimensional data is hard to understand. But is it truly cursed? To help you build better intuition for high-dimensional datasets, @lekschas.de and I developed dtour, a visualization tool for smoothly navigating through data projections.

Seth Levine gave a great talk on the art of clustering at ODSC East that really digs in the practical realities of getting the most out of your data. www.youtube.com/watch?v=LtZ3...

People should not sleep on the new EVoC clustering algorithm. It's got me rethinking a lot of my widgets and bulk labelling tricks. youtu.be/1roN51JXuX8

What’s the world searching for right now? Explore the past 24 hours in combined google trends and local news sources from 125 countries. Similar topics cluster together. Focus on the clusters or the outliers. 🐸 👉 trends.datafrosch.fun #dataviz #dataexploration worldnews #UMAP #localmodels

Map versus territory of XKCD colors. github.com/venkatasg/xkcd…

EVoC is a library designed specifically for fast clustering of high dimensional embedding vectors. It can produce high quality clusters extremely efficiently, and requires little to no hyperparameter tuning. Better clustering than UMAP + HDBSCAN; faster clustering than KMeans.

There's a new release of fast-hdbscan out now on PyPI that adds some major new features. - A PLSCAN implementation based on work by Jelmer Bot - Support for diverse metrics via PyNNDescent - A powerful "cannot-link" constraint system for semi-supervised clustering by Richard Hakim

One of the nicest things about Nvidia model releases is that they ship the training data. What does it look like? I sampled 250k examples from 24 datasets in the Nemotron post-training v3 collection and built an interactive Embedding Atlas to explore it.

This week I'll be delivering a talk and leading a tutorial for the Language AI in Space Sciences Workshop in Baltimore, hosted at STScI. So excited to see folks from all disciplines join the conversation -- including astronomers, engineers, computer scientists, librarians, and linguists.

Han Jiao (ex jina Ai ceo) built mlx apple silicon clustering libs for umap, tsne etc github.com/hanxiao/mlx-...

playing around with umap-js today.

Excited to share a new #interactive #dataviz, putting 50 years of #neuroscience research on the map: The State of Neuroscience 2025 stateofneuroscience.thetransmitter.org #StateOfNeuro

Any sufficiently large k-nn is indistinguishable from magic 🧙‍♂️

become unclusterable

I made a map of 3.4 million Bluesky users - see if you can find yourself! bluesky-map.theo.io I've seen some similar projects, but IMO this seems to better capture some of the fine-grained detail

Our new pre-print shows how unsupervised clustering methods can identify biologically meaningful differences in early vocal production, with no human feedback. @antorrisi.bsky.social has led this interdisciplinary collaboration based on computational methods + #chicks 🐣 arxiv.org/abs/2601.12203

here's a fun side project i've been working on: i compiled a joint text<>audio embedding model to a fast coreml pipeline, and built a very fast (~400ms for 50k samples, can scale to millions) UMAP dimensionality reduction GPU impl in mlx. using it to browse music libraries and do sample sim search

Xiaobin Li, Run Zhang: Understanding and Improving UMAP with Geometric and Topological Priors: The JORC-UMAP Algorithm https://arxiv.org/abs/2601.16552 https://arxiv.org/pdf/2601.16552 https://arxiv.org/html/2601.16552

I miss the days where you'd see blogposts with clever analyses on datasets, maths and data science tricks. That's why, as an experiment, we're starting a new moderated subreddit. People can share/promote their notebooks and you can use RSS to subscribe. Please join and share!

UMAP connectivity plots of 3,627 chess openings from the @lichess.org datasets (huggingface.co/datasets/Lic...)

I think it's important to note though that in spite of those incentives, the direction of the last two years has been more fungibility, *not* lock-in. And open source is the wrong fight here: when lock-in comes it will look more like the lock-in that Amazon or Uber have than Microsoft Office…

New preprint! Have you ever wondered, what are these fuzzy simplicial sets, the theoretical framework behind e.g. UMAP? Here we show that you may simply see them as marginal distributions over simplicial sets. This provides a generative model for UMAP. (1/2) arxiv.org/abs/2512.03899

Space DJ turns genre embeddings into a playable galaxy—pilot a ship, the music follows. 🚀 Key stats 768→128 PCA compression; 3D UMAP projection; three.js rendering; autopilot drift; high‑dim neighbors surfacing hidden similarities.

via the magic of laion_clap embeddings and umap, my live coding thingy has a sample browser at last!

I made this annotated scatter plot of 1 million FineWeb-Edu documents for @sashamtl.bsky.social's new TED talk.

Also really love how organic the plot looks with "inferno" (left) and "viridis" (right).

Map of the internet: 1.3M nodes (BGP)

The video of my talk at SciPy on DataMapPlot is up at last. If you make t-SNE or UMAP plots the talk provides some guidance on how to make plots most effective, and introduces a library to help make that easier. www.youtube.com/watch?v=-iBh...

Despite the gutting of the National Center for Educational Statistics, the dept of Ed *did* manage to release 2024 college major counts in the usual format, so I can run it through the same code I do every year. First off, the change since peak of the largest fields -- another year of drops.

I'm very much a learner, but you're maybe asking if aspects of matrix factorisation approaches to dimensionality reduction apply here. But LocalMAP is a KNN approach, with a matrix factorisation initialisation. h/t @lelandmcinnes.bsky.social for his attempts to describe these youtu.be/9iol3Lk6kyU

📢 Save the date! Join us for the next @ellis.eu x UniReps Speaker Series! 📅 27th August – 16:00 CEST 📍https://ethz.zoom.us/j/66426188160 🎙️ Speakers: Keynote by @lelandmcinnes.bsky.social & Flash Talk by Yu (Demi) Qin 🔔 Stay updated by joining our Google group: groups.google.com/u/2/g/ellis-...

🚀 We've just open-sourced Embedding Atlas – a tool for exploring large embedding spaces through rich, interactive visualizations 📊.

Meteoroid stream identification with HDBSCAN unsupervised clustering algorithm. Eloy Peña-Asensio et. al. https://arxiv.org/abs/2507.01501

Ever wanted to pan through the latent🌌 space of TikTok videos? Made using the amazing toponymy and datamapplot from @lelandmcinnes.bsky.social and data from mine and @jurgenpfeffer.bsky.social 's first complete TikTok slice. link below

🎤 Speaker Spotlight: Leland McInnes Join Leland at #SciPy2025 for his talk "DataMapPlot: Rich Tools for UMAP Visualizations." 📊 Discover powerful new ways to explore high-dimensional data! 🔗 scipy2025.scipy.org

Explore Wikipedia through a data map. Pages are grouped by semantic similarity, for topic clusters. Hover to see details, zoom to explore more fine-grained topics, click to go to a page. Search by page name to find interesting starting points for exploration. lmcinnes.github.io/datamapplot_...

I'll be giving a talk about DataMapPlot for visualizing data maps at Scipy this year. I would love to meet potential users and chat about where to go next. cfp.scipy.org/scipy2025/ta...

I also updated the ArXiv data map example to make use of new features in datamapplot. lmcinnes.github.io/datamapplot_... You can tweak parameters and build your own version: gist.github.com/lmcinnes/e11...

OMG I am so glad someone finally did this. Thank you 🙏 @lelandmcinnes.bsky.social This will now consume hours and hours of my time. lmcinnes.github.io/datamapplot_...

I also updated the ArXiv data map example to make use of new features in datamapplot. lmcinnes.github.io/datamapplot_... You can tweak parameters and build your own version: gist.github.com/lmcinnes/e11...

Great idea. Did no one think of this before?

Explore Wikipedia through a data map. Pages are grouped by semantic similarity, for topic clusters. Hover to see details, zoom to explore more fine-grained topics, click to go to a page. Search by page name to find interesting starting points for exploration. lmcinnes.github.io/datamapplot_...

🔥 Meet our Keynote Speakers for #SciPy2025! Dr Malvika Sharan, co-Director of Open Life Science (OLS) and a senior researcher at The Alan Turing Institute will be sharing with us her expertise at our favorite conference. You can't miss her ➡️ hubs.la/Q03sdlsb0

🔥 Meet our Keynote Speakers for #SciPy2025! Hon. Dr. Kathryn D. Huff 🇺🇸, nuclear engineer, policy leader, and former Assistant Secretary for the Office of Nuclear Energy will be joining us in Tacoma! 🙌 Don't miss her talk, grab your ticket now: hubs.la/Q03sdlsb0

Nature Reviews Methods Primers: Uniform manifold approximation and projection (UMAP) www.nature.com/articles/s43... 🧬🖥️🧪 read free: rdcu.be/d0YZT

Our latest paper is out: peerj.com/articles/cs-.... We added functionality to the #HDBSCAN clustering algorithm to also detect branches hdbscan.readthedocs.io/en/latest/ho... #eda #datavis #clustering