Toots for PyDataParis@mastodon.social account

Written by PyData Paris on 2025-01-30 at 10:09

🔍 Building scalable ETL pipelines with Dask is easier than ever. P. Hoefler discusses leveraging Dask for large-scale data processing, highlighting recent improvements in shuffling algorithms and query optimization. Learn how Dask can outperform Spark in cost and performance.

https://youtu.be/oZM4k_vVzMM

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-28 at 08:18

🔬 Spatial Transcriptomics, named the method of the year by Nature in 2020, offers remarkable visuals of gene expression across tissues and organs. Inácio Medeiros presents the Squidpy library for analyzing and visualizing spatial molecular data.

👀 Watch the talk: https://youtu.be/qsYHuNcitzA

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-23 at 11:17

🔍 Collaborative editing in Jupyter now allows real-time document modification, similar to Google Docs. This is powered by Conflict-free Replicated Data Types (CRDTs), enabling decentralized data distribution. In this talk, David Brochart explores new possibilities with CRDTs, including a suggestion system, AI collaboration, and full notebook state recovery. Learn more about how CRDTs are redefining Jupyter's architecture. Watch now: https://youtu.be/VXXLnmGqAO4

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-21 at 12:41

🔍 Evaluating RAG systems is crucial yet challenging. Nour El Mawass and Maria Knorps dive into the stability and usefulness of RAG evaluation libraries. They explore the impact of evaluator LLMs, query reformulation, and dataset characteristics on performance. This talk offers insights into selecting and interpreting LLM-based evaluations effectively.

🎥 https://youtu.be/yrWmK3pXrKk

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-14 at 12:53

🔔 "Would you rely on ChatGPT to dial 911?" Nicolas Guenon des Mesnards explores balancing determinism and probabilism in production ML systems. Learn how to enhance robustness and controllability in ML models. 🔍

Watch now: https://youtu.be/UUNaRYvEJkY

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-09 at 15:35

📊 Propensity Score Matching (PSM) has significant flaws. Alexandre Abraham introduces A2A, a novel metric to enhance PSM reliability.

🔍 Watch now: https://youtu.be/y75GbnHizaQ

=> More informations about this toot | View the thread

Written by PyData Paris on 2025-01-07 at 14:09

🔬 Processing medical images at scale on the cloud is crucial for advancing oncology. Guillaume Desforges discusses the challenges and solutions in handling large Whole Slide Images for cancer detection. Learn how cloud-based solutions and efficient data processing can empower medical research. 🌐

Watch the full talk: https://youtu.be/GHTPLJpzgtM

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-19 at 10:15

🎯 Adaptive prediction intervals are crucial for practitioners involved in decision-making. Learn about state-of-the-art algorithms for constructing these intervals. 🔗 Watch Andro Sabashvili's talk here: https://youtu.be/t56zcfj51aU

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-17 at 09:39

📊 Unpacking business metrics to explain their evolution is a common challenge for data analysts. Max Halford's talk at PyData Paris introduces a robust framework for decomposing metrics, offering a clear methodology and an open-source Python tool. This approach helps attribute changes in metrics to specific drivers, enhancing stakeholder trust and understanding. 🔍

Watch the full talk here: https://youtu.be/kiWW4Oty8kY

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-12 at 12:56

📈 "Half of my marketing budget is wasted, but I don't know which half." This century-old quote still rings true today. Emanuele Fabbiani discusses how Bayesian models can optimise marketing spend, using WeRoad's case study. Learn how to maximise ROI with actionable insights. 🔗 Watch now: https://youtu.be/GYmE_VcAeJA

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-10 at 13:09

🔒 The EU's Cyber Resilience Act is set to impact the software industry globally. @cheukting_ho discusses its implications, timeline, and new responsibilities for developers and maintainers. Learn how the Python Software Foundation is preparing and how you can too.

📺 Watch now: https://youtu.be/KkUo8bD-0cE

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-05 at 11:01

💡 Open-source AI is revolutionizing the field, offering privacy, customization, and cost-efficiency. @mervenoyan from

@huggingface explores why it matters and how to get started. 🔗 Watch now: https://youtu.be/-Mv1KGr2FKs

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-12-03 at 14:55

🔍 Long-term knowledge acquisition requires repetition. Nicolas M. Thiéry introduces Jupylates, a Jupyter extension for spaced repetition in teaching. This tool integrates repetitions directly into Jupyter courses, personalized on learning history.

📚 Watch the full talk: https://youtu.be/4WDVm4PcHig

[#]PyData #PyDataParis

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-11-26 at 13:39

📢 The videos of our 2024 conference are out !

Go get a look and enjoy hours of interesting talks. 💻

[#]PyData #PyDataParis

https://youtube.com/playlist?list=PLGVZCDnMOq0pKya8gksd00ennKuyoH7v7&si=Y3BM290s-i0E6QK1

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-09-11 at 14:44

Note that the (free) registration to the sprint day is independent of the (non-free) registration for the main PyData Paris conference days.

Registering for the sprint is done on the PyData Paris meetup page:

https://www.meetup.com/pydata-paris/events/303367322/

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-09-11 at 09:54

Why Join?

💡 Contribute to open-source projects, learn from experts, and connect with the community!

🌟 Benefits:

Make open source better!

Code alongside experts,

Learn new dev skills, tools and tricks,

Improve your GitHub profile,

Network with the community.

💻 All levels welcome! Sprints are free.

🔗 Interested in leading a sprint? Submit your project here: https://forms.gle/wtTgjXJpVCEhWdjj7

=> More informations about this toot | View the thread

Written by PyData Paris on 2024-09-11 at 09:53

📢 Join us for PyData Paris Sprint Day on Sept 27, 2024! 📢

📅 Date: Sept 27, 2024

🕒 Time: 9:30 am - 5:00 pm

📍 Venue: Carrefour du Numérique, Cité des Sciences et de l’Industrie

What is a Coding Sprint?

A coding sprint is an informal event where developers collaborate to contribute to open-source projects. It's a great way to learn, network, and make a meaningful impact on the tools you use every day.

[#]PyData #OpenSource #DataScience #Python

=> More informations about this toot | View the thread