iiinova — where your digital self keeps evolving

Florian Huber @me-datapoint.bsky.social · 25/09/2026

Die politischen Meldungen der letzten Tage kann ich wirklich bald nur noch in Form von Satire ertragen.

Small furious dog (CDU) barking at harmless dog (LINKE) while evil combat dog is waiting on the right (AFD).

Florian Huber @me-datapoint.bsky.social · 14/09/2026

How to intuitively explore abstract, complex data using interactive #dataviz? --> was great to show people some of our projects at the "Nacht der Wissenschaft" @zdd-hsd.bsky.social

001

Florian Huber @me-datapoint.bsky.social · 07/09/2026

Good #Dataviz is very hard to master because even small things can have a big effect on how a visualization is perceived. Today I'd like to thank @taz.de for consistently coloring German election plots correctly. Doesn't make the results any better, but at least it removes the false facade.

Election results from Sachsen-Anhalt (2026) as shown in the taz with AFD results in proper BROWN color and not their harmlessly-looking blue. Taken from: https://taz.de/Landtagswahl-in-Sachsen-Anhalt/!6188290/

140

Florian Huber @me-datapoint.bsky.social · 09/07/2026

Here's v1.0 of my data science script 🚀 Last week of my introduction to #datascience course this semester. I used this as another chance to reiterate, improve, expand my teaching materials. The entire course material --> florian-huber.github.io/data_science... #oer #openscience #python #teaching

2165

Florian Huber @me-datapoint.bsky.social · 30/06/2026

7/ Finally, we built a helper Python library: #chemap for handling all the different fingerprint types and variants (including unfolded and frequency-folded), and also allows fast fingerprint and score computations. Contributions/Feedback are very welcome! --> github.com/matchms/chemap

010

Florian Huber @me-datapoint.bsky.social · 30/06/2026

5/ (3) There is no clear "winner", so no fingerprint that does better on all or our benchmarking tasks. When in doubt, however, Morgan and FCFP count fingerprints, ideally with a larger radius, are good starting points. They also correlate well with graph-based MCES similarities.

Spearman correlation plot between various fingerprint-based similarities and rascalMCES

110

Florian Huber @me-datapoint.bsky.social · 30/06/2026

4/ (2) High-bit-number fingerprints (e.g., RDKit, MAP4) tend to strongly favor larger over smaller compounds and produce arbitrary high-similarity predictions. This can be fixed by simply using unfolded variants! We also proposed and explored "frequency-folded" fingerprints for ML tasks.

Tanimoto deviation between unfolded and folded.

110

Florian Huber @me-datapoint.bsky.social · 30/06/2026

3/ (1) Counting substructure or bit occurrences drastically increases specificity. For binary fingerprints, we typically observed a high fraction of identical fingerprints for notably differing compounds. Also, chemical class prediction and neighborhood consistency improve when using counts.

Fingerprint duplicates highlight the various specificity of fingerprints.

110

Florian Huber @me-datapoint.bsky.social · 19/05/2026

Another round of edits on my "Hands-on Introduction to Data Science" book. Finally managed to add a bit more explanatory text and figures on UMAP in the dimensionality reduction chapter. --> florian-huber.github.io/data_science... #OpenScience #teaching #python #datascience #oer

nearest neighbor graph computed for spiral shaped data point distribution.

130

Florian Huber @me-datapoint.bsky.social · 01/04/2026

Working on another iteration of my #DataScience introductory course; Revising/expanding materials and new figures and illustrations. Here is a new sketch to display the rich data science skills landscape (not meant to be exhaustive). --> The book: florian-huber.github.io/data_science...

Illustration of the rich data science skills landscape (CCBY Florian Huber)

031

Florian Huber @me-datapoint.bsky.social · 27/02/2026

Just updated our preprint on benchmarking molecular fingerprints! --> www.biorxiv.org/content/10.1... Some key points - Count fingerprints should be the default (not binary!) - Unfolded fingerprints are often worth it. - Larger radius Morgan or FCFP fingerprints are good first bets.

MST graph comparing the top-10 ranking overlap between many commonly used fingerprint types and variants.

Four UMAP visualizations of chemical space based on 700k fingerprints on the biostructures dataset using various molecular fingerprint types.

172

Florian Huber @me-datapoint.bsky.social · 06/10/2025

New #matchms release (0.31)🚀 With functionalities that were on our TODO list for a looooong time: Flash Entropy and BLINK scores! The new "FlashSimilarity" allows computing modified cosine, spectral entropy etc., about 100x faster (or more if you use Linux). #Python #opensource #massspec

162

Florian Huber @me-datapoint.bsky.social · 26/06/2025

@jorainer.bsky.social and @philouail.bsky.social gave a great overview of the ecosystem around #RforMassSpectrometry and #XCMS! #MetSoc25 I am super glad they now also provide options to combine with #Python and #matchms (thanks🙏)

1116

Florian Huber @me-datapoint.bsky.social · 25/06/2025

Great keynote by @sneumann.bsky.social at #MetSoc25, strongly advocating for #opensource , data-sharing, and making things interoperable. Glad to also spot #matchms in this universe :)

Slide from presentation of Steffen Neumann

2174

Florian Huber @me-datapoint.bsky.social · 23/06/2025

4/4 We also highlight options for count fingerprints, such as log-counts and IDF weighted counts. The latter can be used to adjust the bit importance to a dataset of your choice. An example use-case are chemical space visualizations. Preprint: www.biorxiv.org/content/10.1...

Chemical Space Visualizations using UMAP and various molecular fingerprints.

021

Florian Huber @me-datapoint.bsky.social · 23/06/2025

3/4 A huge issue is bit collisions. Fingerprints with a high bit occupation (RDKit, MAP4) often lead to (1) arbitrary misinterpretations, (2) shifts to high Tanimoto scores, (3) very different handling of small and large molecules. --> Consider using sparse fingerprints! --> Morgan >> MAP4 / RDKit

121

Florian Huber @me-datapoint.bsky.social · 23/06/2025

2/4 We focused on weaknesses of the fingerprints. Many show frequent duplicates, so same fingerprint for different compounds. Most problematic: this can include *very* different compounds ending up with identical fingerprints. - MAP4 >> Morgan-type >> daylight - count >> binary #cheminformatics

Benchmarking plot on fingerprint duplications.

101

Florian Huber @me-datapoint.bsky.social · 23/06/2025

New preprint out! 1/4 @julianpollmann.bsky.social and I went down several rabbit holes to assess some commonly used molecular fingerprints. Bottom line: For large datasets, make an effort to select suitable settings. "We used Tanimoto" is not good enough. --> www.biorxiv.org/content/10.1...

Sketch of count/binary fingerprints and weighing options.

151

Florian Huber @me-datapoint.bsky.social · 23/06/2025

Good start for me at #metabolomics2025 with a hands-on workshop on MS2LDA by Jonas Dietrich, Rosina Torres Ortega and @jjjvanderhooft.bsky.social.

053

Florian Huber @me-datapoint.bsky.social · 22/06/2025

Went by train to #Prague for #metabolomics2025. These are the kind of moments that remind me how great the European project is. No border controls, no visas. Just a train following a river to the neighboring country.

Elbe river seen from a train somewhere after Dresden.

0150

Florian Huber @me-datapoint.bsky.social · 17/06/2025

Da kann der Bürgermeister @duesseldorf.bsky.social noch so oft die "Fahrradhauptstadt" (🥹🤭😭) herbeibeschwören... es braucht dann doch ein bisschen mehr als ein paar Kleckse Farbe. #Düsseldorf weiterhin konstant bei 4- im #ADFC Klimatest. Läuft. @adfcnrw.bsky.social @adfc-duesseldorf.de

Screenshot vom ADFC Fahrradklima-Test 2024 für Düsseldorf.

100

Florian Huber @me-datapoint.bsky.social · 06/06/2025

When you prepare lesson material while being hungry... (added some text edits and more sketches/figures to the NLP chapters of the "Hands-on Introduction to #DataScience with #Python" textbook) florian-huber.github.io/data_science... #OpenScience #Teaching #CCBY

040

Florian Huber @me-datapoint.bsky.social · 14/05/2025

New release of my "Hands-on Introduction to Data Science with Python" textbook! Contains many text edits and figure updates. For instance, in the sections on Clustering and Machine Learning. All fully #opensource and #openaccess. Figures are #CCBY. --> florian-huber.github.io/data_science...

1236

Florian Huber @me-datapoint.bsky.social · 25/03/2025

Here is a short blog post on the typical data science workflow (if something like that even exists). Happy to take any feedback or suggestions. --> medium.com/@f.huber/wat... #DataScience #Python #Teaching

030

Florian Huber @me-datapoint.bsky.social · 12/03/2025

#matchms workshop in full swing! Great fun to work with this fantastic group of people on improving, expanding, applying matchms for handling #massspec data in #Python.

1102

Florian Huber @me-datapoint.bsky.social · 03/12/2024

It took us quite a while... not easy with most authors undergoing career switches (me, too). But finally the last bit of my postdoc work with @marileend.bsky.social is now published in @biophysj.bsky.social. #microtubules with EB3 growing against rigid barriers. --> www.cell.com/biophysj/ful...

Microtubules with EB3 growing against rigid barriers.

1122

Florian Huber @me-datapoint.bsky.social · 22/11/2024

Enjoyed exploring the beautiful map of #bluesky (from 2024-11-07) from @syntacrobat.xyz! --> aurora.ndimensional.xyz

UMAP plot of all bluesky accounts of 2024-11-07 made with https://aurora.ndimensional.xyz/.

072

Florian Huber @me-datapoint.bsky.social · 20/11/2024

Delighted that a project that took shape within a #machinelearning workshop at the @esciencecenter.bsky.social and which was initiated and led by #RozaKamioglu and #DisaSauter led to a data analysis of human #laughter sound --> royalsocietypublishing.org/doi/10.1098/... #OpenAccess #DataScience

Sketch of analysis framework for laughter context.

270

Florian Huber @me-datapoint.bsky.social · 15/11/2024

I continue to work on #opensource versions of my various teaching materials. Here is the first complete draft for my Python Introduction (for the moment only in German, but English version is on the Todo-list) --> florian-huber.github.io/python-intro... #DataScience #teaching

Cover for the German Programming with Python for Data Science Course.

3164

Florian Huber @me-datapoint.bsky.social · 14/10/2024

New version of my (renamed) textbook: "Hands-on Introduction to Data Science using Python" 🎉 The content is now mostly complete. Text and figures will undergo further polishing. --> florian-huber.github.io/data_science... #datascience #opensource #teaching

0130