PyData NYC

Miss PyData NYC this year? Check out the replay: Eliminate Hallucinations with Context Filter Guardrails

I wanted to share an innovative presentation technique I developed that helps me deliver deeply personal talks with consistency and emotional precision. I create memory anchors using color-coded emotional acronyms for each slide. Here's how it works: For each slide, I first write out my narrative verbatim. I then feed this text to an AI, asking it to create an acronym where each letter corresponds to a key sentence while capturing the slide's emotional essence. The order of letters is crucial - for example, in 'RISE', the 'R' corresponds to my opening sentence, and the 'E' maps to my closing point. These acronyms become visual anchors, appearing as colored circles in the corners of my slides. The circle's color represents the emotional tone I need to convey. For instance, in a recent talk about my life journey: - A green 'RISE' circle reminds me to convey ambitious optimism when discussing my path to earning a Ph.D - A pink 'LOVE' circle signals warmth and tenderness when sharing memories of my grandmother - A black 'DEPTH' circle guides me to express profound gravity when discussing my decade-long battle with bipolar depression This system helps me maintain emotional authenticity while ensuring I don't miss crucial narrative points. It's especially valuable for talks dealing with sensitive personal experiences where both content and emotional delivery need to be precisely calibrated. Would love to hear if others have developed similar techniques for managing emotional delivery in presentations. #PresentationTechniques #PublicSpeaking "PyData NYC

Thank you to everyone who joined us for our final PyData NYC meetup of the year! A special thank you to our speakers, Milan Janosov, Kelly Abuelsaad, and Thanos Tatsios, for taking the time to share your expertise with us! We wanted to mention that this year has been incredible because of your support. We are so grateful to have such an amazing community, and we look forward to seeing new and familiar faces in the year ahead. Here's to more learning, sharing, and growing together in 2025!

?? DataFrame dilemma? Dharhas Pothina's PyData NYC talk is your ultimate guide! Dive deep into pandas, Polars, Dask & more. Learn which tool fits YOUR data workflow. Watch now: ?? https://lnkd.in/dt3v9UAU #DataScience #PyData

Everything Everywhere All at Once ?????? The past few weeks have been a whirlwind—bouncing between the concrete jungle of PyData NYC (shoutout to the amazing open-source community!) and the broader context of the U.S. elections. I’ve barely had a moment to reflect, but one thing remains clear: the power of community and collaboration has never felt more important. ??? As we approach the end of 2024, I’m inviting you to join NumFOCUS in championing open-source innovation. Whether it’s by pledging your support to our End-of-Year Campaign or joining us for PyData Global (Dec 3-5), there’s never been a better time to connect, learn, and grow together. ?? Because when we support open source, we’re building tools—and communities—that empower everyone. ?? Special shout-out to a few folks who made PyData NYC not just a moment, but a good vibe: Leah Silen Arliss Collins Nolan Fortman Jonathan Starr Tomara Youngblood Kelby Lorenz Tim Bonnemann Katrina Riehl James Powell Christa Ruggiero John Carney Dawn Gibson Wages Na'im Tyson Mars Lee Christopher Fonnesbeck Lawrence Gray, Ph.D. #NumFOCUSImpact #fuelthefuture #opencodeforgood #opencode4good #globaldev

Thrilled to be back at PyData NYC second year in a row! ?? It was great coming back to the conference last week in NYC for a 3-day in-person event for the international community of data practitioners to share ideas and learn from each other. With a series of informative workshops and sessions, the ones that I found very interesting are: ?? 1. Nidhin Pattaniyil's hands-on workshop on building RAG in Python on YouTube videos. It covered various topics of information chunking, Embedding Generation, and Retrieval and was fascinating! 2. Allison Wang's session on how Apache Arrow improves PySpark's performance covering features of Pandas and Arrow UDFs. 3. Santosh Borse's session on Preparing data for LLM Training and introduction to Data Prep Kit on how it has navigated the journey of improving data quality & governance. As last year, it was great to meet and network with many amazing data enthusiasts from different parts of the globe and learn from their experience! Anyone interested in learning more about the materials shared? Feel free to DM ?? #PyDataNYC #dataconference #dataengineering

Two weeks ago I was at PyData Global #PyDataNYC conference and I have still yet to recover from its awesomeness (the work week coming back might have something to do with it ??). Here are some of my highlights: ?? it was an great to meet Leah Silen and the rest of NumFOCUS crowd (Jonathan Starr, Tomara Youngblood, ...), along with PyData London chair John Carney, and talk about potential future activities in Italy ???? workshop day started with a nice introduction on how to build multi-modal apis by Hugo Bowne-Anderson (check out his podcasts if you haven't yet, so much great material) ?? it was lovely to connect with people from the Posit PBC crowd: Michael Chow (learned about reactables, looking forward to use them), Isabel Zimmerman (Positron looks really nice, the parquet explorer feature sold me right away), Carlos Scheidegger (thanks for your quick walkthrough on Quarto's codebase, looking for slides' css) ??I was so happy to discover the Milan connection ?? with Jessi Shamis who leads the design of one of my favorite tools, Streamlit (looking forward to meet again on the other side of the pond) ??? Lawrence Gray, Ph.D. keynote on stories and values from life and open source was both entertaining and emotional (especially the second part on the topic of mental health) and it will be one of the strongest moments I will remember of this conference ?? volunteers are fundamental for any conference, I am happy to have a new friend that was a volunteer at the conference, Mars Lee (link below on the awesome "how to contribute to numpy" comic!) ?? the most charming talk was defintely the one by Wojciech Matejuk on a transform based approach to generate music (that captures the subtle beauty of humane performances) ?? a recurring thread on many talks was the trend towards more dense compute, the topic of Benjamin Zaitlen's keynote, and in general a number of tools that in one way or another are innovating the Big Data space: Bryce Adelstein Lelbach talk was a great introduce on how to program a GPU in pure Python with CudaPy, Dask is now faster than Spark (thanks Patrick H?fler for giving me a summary of the 3 reasons why), there is a new Rust-based tool in town for streaming workloads called Bytewax (thanks to Federico Dolce I was not the only Italian at the conference, also Zander Matheson you definitely had the best gadgets, my daughter loved the soft bee keychain??), Ritchie Vink shared how Polars will have a GPU and distributed computing backends, ... ?? ... I could go on, but let me stop by mentioning that it was a pleasure and an honor to share my thoughts about the importance of domain expertise in Data Science (link to slides below) ???? A final thanks to Agile Lab, I am happy and proud that participating and sharing with the wider community is part of our values and our way of working.

Great insight!

Had a great PyData NYC conference last week. There were 5 talks/tutorials with Polars as topic, which means we are doing great! Jeroen Janssens and Thijs Nieuwdorp, the authors of the upcoming O'Reilly book "Python Polars: The Definitive Guide" were presenting there as well. And I finally met ?? Matt Harrison IRL. He is known for many books on pandas, Polars and more. And I got a hardcopy of his "Effective Polars" book, making our Polars library count increment to 2. ?? And last but not least Ollie Dapper flew by to join his first Pydata, making it feel a bit like Xomnia fun trip.

?? Last week, PyData NYC inspired and energized us! Huge thanks to NumFOCUS for organizing such a fantastic event. The Bytewax team has returned home, taken a deep breath, and is now back in action. From the booth to the talks, it was all about connecting with the community. We loved hearing your thoughts, answering your questions, and sharing our approach to real-time data processing. ?? ?? Zander Matheson’s session focused on how Bytewax combines Python’s flexibility with Rust’s speed for real-time streaming. He shared practical tips on using PyO3 to optimize performance, integrating Rust with Python, and packaging with Maturin, offering a clear roadmap for high-efficiency workflows. If you’re interested, you can check out the slides [the link is in the comments]. We’ll also be sharing the recording soon for those who missed it. ?? Stay tuned!