srvanderplas (@srvanderplas@datavis.social)
@sennoma understanding which of #rstats or #python should be used for different tasks is a better option. #Seaborn is ok, but #ggplot2 is in another class. But there are lots of text mining methods and some machine learning things that are better developed in python. Data science is multilingual and focusing on concepts and specific implementations is more important.
Show Original Post
giggls (@giggls@karlsruhe-social.de)
Hey #Openstreetmap people. Is there a simple way to check the correct function of a slippymap e.g. using #python and #selenium with tools like #icinga or #checkmk?
Show Original Post
darc (@darc@social.darc.de)
NASA-Cybersec trifft Amateurfunk: Deep Learning gegen das Rauschen 🚀📻
Jimi Sanchez (NASA Cybersecurity Manager) hat ein spannendes Paper veröffentlicht, das zeigt, wie wir die unberechenbare Physik der Funkwellen mit neuronalen Netzen bändigen können.
Jimi ist selbst leidenschaftlicher Funkamateur (KJ1MBO) und nutzt seine Expertise aus der Absicherung von NASA-Cloud-Infrastrukturen, um die Signalklassifizierung zu revolutionieren.
Die harten Fakten:
Hybrid-Ansatz: Einsatz von CNNs für die visuelle Mustererkennung in Spektrogrammen und RNNs für die Analyse zeitlicher Signalabfolgen.
Robustheit: Durch „Data Augmentation“ (künstliche Signalverschlechterung) wurde das System auf die chaotischen Realitäten der Ionosphäre trainiert.
Performance: Signifikante Steigerung des SNR (Signal-to-Noise Ratio) und Reduzierung der Bitfehlerrate (BER) bei digitalen Betriebsarten.
Statt starrer Filter nutzt Jimi lernende Systeme, die sich in Echtzeit anpassen – ein echter Fortschritt für #OpenScience und die #HamRadio Community.
👉 Das vollständige Paper (PDF, Englisch) gibt es hier: https://arxiv.org/pdf/2402.17771
#AmateurFunk #NASA #Cybersecurity #MachineLearning #DeepLearning #HamRadio #SDR #SignalProcessing #Python #Arxiv #KJ1MBO
Show Original Post
rzeta0 (@rzeta0@mathstodon.xyz)
Q: what's the best* python implementation of the grammar of graphics ?
* means popular, established, sustainable dev, publication quality plots, worth investing time learning over others - but doesn't mean most featureful
tagging ggplot2 because that community is likely to know
#python #ggplot2 #dataviz #datascience
Show Original Post
wraptile (@wraptile@fosstodon.org)
Chancy is a brilliant modern alternative to Celery through postgresql. Quite minimalistic but really powerful and a well designed task queue framework. Check it out:
https://github.com/tktech/chancy
Show Original Post
DSLCvids (@DSLCvids@fosstodon.org)
Recent @DSLC club meetings:
Advanced R: Expressions https://youtu.be/8HmfuourNyI #RStats
From the @DSLC ​chives:
ISLP: Multiple Testing https://youtu.be/sOgUOTwF4ek #machineLearning #pydata #python
ISLR: Multiple Testing Part 1 https://youtu.be/iwd20sDYUbY #RStats
Advanced R: Measuring performance https://youtu.be/4hngR1c9oP4 #RStats
Support the Data Science Learning Community at https://patreon.com/DSLC
Show Original Post
hugovk (@hugovk@mastodon.social)
Some new preliminary numbers from Ken Jin for the JIT in Python 3.15:
"On AArch64 macOS, the JIT has a 7-8% speedup over the tail calling interpreter with all optimizations enabled."
https://docs.python.org/3.15/whatsnew/3.15.html#whatsnew315-jit
https://github.com/python/cpython/commit/d9c1235db44100b16355a347ca9e304df48411a7
This is in addition to 3-4% over the standard CPython interpreter on x86-64 Linux.
See also https://doesjitgobrrr.com
#Python #Python315 #JIT #CPython
Show Original Post
ENTERCO (@ENTERCO@mastodon.social)
¡Impulsa tu carrera en tech! InscrÃbete GRATIS al certificado profesional de Python de Microsoft y aprende
desde cero. 🚀
#Python #MicrosoftCertificación
Show Original Post
spacehobo (@spacehobo@teh.entar.net)
I am "#Python is the batteries-included language, unlike Perl or TCL!" years old, and that has aged like a prawn cocktail. The standard library is chok full of awkward "un-#Pythonic" legacy code that only exists because changing it would break backward-compatibility.
But Python 3.0 *gleefully* broke backward-compatibility! They broke `print "Hello, world."` which is the most-written Python code in history!
So for my money, Python 4.0 should be a release that gets rid of boat anchors like HTTPlib, and pushes that stuff into a `from deadbatteries import ...` compatibility library for the folks who absolutely need it.
You could do statistical analysis of `requirements.txt` files out on the Internet and see what replaces-the-horrible-stdlib-version packages are imported the most. You can then mainline the winning replacements, or figure out what APIs or features everyone is reaching for when they import some new dependency and reimplement that. It doesn't matter if you NIH it, re-do the API on existing library code, or promote a third-party project to Python core library status.
Python's dependency resolution tooling is not just a mess: it's a dozen independently-maintained messes that compound the problem. I want to go back to being able to actually rely on the standard library. I do my best now, but the level of jank is only getting worse as the years drag on, and that prawn cocktail is absolutely pongin'.
Show Original Post
NatuRandom (@NatuRandom@mastodon.social)
Thamnophis proximus rubrilineatus
PH Alejandro Santillana
Texas, United States
2017-06-17 15:21:00 UTC
https://www.inaturalist.org/observations/6698340
#iNaturalist #Nature #Wild #Python #FedivEarth #Environment
Show Original Post
isazajuancarlos (@isazajuancarlos@mastodon.social)
Potencie su negocio con soluciones Python de Xiliux. Desde MVPs hasta integración de IA, ofrecemos automatización y scraping a partir de $200 USD. Transforme su idea en realidad. #LegalTech #Python #Xiliux
Show Original Post
2026 (@2026@elradioescucha.net)
Software para gestionar frecuencias con SDRSharp
SDR# Frequency Manager un nuevo software desarrollado para facilitar la gestión de listas de frecuencias en SDR#.https://elradioescucha.net/2026/01/07/software-para-gestionar-frecuencias-con-sdrsharp/
Show Original Post