PyMC Labs

???? Happy Thanksgiving from PyMC Labs! ???? Today, we take a moment to reflect on the incredible community that makes PyMC Labs what it is — a space for innovation, collaboration, and growth. We’re thankful for: ? Our brilliant contributors who push the boundaries of Bayesian modeling. ? Users and learners who inspire us with their passion for probabilistic programming. ? Partners and collaborators who help us bring impactful projects to life. ????'???? ???????????????? ?????? ?????? ???????????? ???????? ?????????? ???? ??, ?????? ???????? ???????? ???????????????? ???? ??, ?????? ?????? ???????????????????? ???????? ?????????? ?????? ?????????????? ? May your day be filled with warmth, joy, and maybe even a touch of statistical modeling magic! ?? #Thanksgiving #Gratitude #PyMC #BayesianModeling

?????????????? ?????????? ??????????: ????’???? ???? Bluesky Social! ??? ? ?? You can now follow us at @pymc-labs.bsky.social(https://lnkd.in/gaidxaZ3). We’ll be sharing updates, and thoughts, and diving into some great conversations over there. Bluesky Social is growing fast, surpassing 21 million users, leaving all prior expectations in the dust?? and showing no signs of slowing down. ??We look forward to joining the conversation and getting into insightful Bayesian discussions with everyone. See you there! ?? #BlueSky #PyMCLabs #PyMCMarketing #CausalPy

?? The PyData NYC talk by Christian Luhmann on ????????-???????????????????is now live on Youtube! ? Whether you're exploring how to use data for better business decisions or curious about media mix models or customer lifetime value, this session is packed with insights. ?? Catch the replay and see how Bayesian modelling simplifies complex analytics! ?? https://lnkd.in/gvs9bxk7 #PyMCMarketing #DataScience #BayesianModeling

?? Finding Meaning in Models: The Balance Between Simplicity, Complexity, and Causality At the recent #Causal Data Science Meeting, I joined Benjamin Vincent, DPhil to discuss one of the most fundamental challenges in Media Mix Modeling (#MMM): How do we build models that don’t just predict outcomes but truly help us understand the why behind them? In our talk, we showcased how PyMC bridges two transformative approaches: causal reasoning and Bayesian frameworks. Together, they form a powerful toolkit for causal reasoning, discovery, and identification, giving practitioners a powerful way to bring structure and meaning to their Media Mix Models (MMM). Here’s the key insight: ?? A simple model might miss key relationships, leading to oversights. ?? A complex model can overwhelm with unnecessary assumptions, risking lack of clarity and correlational bias. The right model isn’t about complexity or simplicity—it’s about causality. With the right causal structure, even a simple model can deliver insights grounded in reality. Without it, even the most sophisticated model is vulnerable to misdirection. My smart collegue Dr. Juan Camilo Orduz aptly put it: > “MMM without a causal DAG is doomed to fail!.” (https://lnkd.in/dhQwxf9X) What’s crucial here is that causality and Bayesian frameworks complement each other perfectly. Bayesian methods provide the flexibility to encode beliefs, test hypotheses, and adapt models as new data emerges. Causality, in turn, brings structure and focus, ensuring the assumptions (beliefs) behind the math reflect reality, allowing to test it. That’s why the bayesian causal identification process we introduced is so exciting. It helps users: ?? Spot and correct when their hypotheses or DAGs don’t align with the data. ?? Identify adjustment sets based on sound causal theory. ?? Bring transparency to the hidden structures within their ecosystem. Causal models don’t just provide better predictions—they provide a framework for understanding. And in the fast-paced world of marketing, understanding is what turns data into decisions. With PyMC-Marketing, we hope to bring clarity to the art of marketing by making the math more meaningful and the assumptions more visible. ??Slides: https://lnkd.in/dFyE98QW ??Notebook: https://lnkd.in/dg3G6Apm #CausalInference #MMM #MarketingScience #PyMC #DataScience

?? Great fun presenting at Python Ireland PyCON 2024 this weekend! ?? I spoke about the managing uncertainty in causal inference using PyMC Labs's and Benjamin Vincent, DPhil's CausalPy package. Slides can be found here: https://lnkd.in/e7WafepT

?????????????? ???? ?? ?????? ?????????? ?????????????? ??????????? ?????????? ???????????????? ?????? ?????? ???? ?????? ???????? “????????????” ???? ?????????????? ?????? ?????? ?????????? ?????????? ?? ?? I wanted to try out the new CausalPy package for causal inference and was brainstorming interesting research questions to apply it to. My inspiration came from a recent wave of LI posts in my feed, all sharing a chart showing a drop in Stack Overflow traffic after ChatGPT’s release in November 2022. ?? Given that many studies - including one we conducted internally at Sanofi - show that one of the most frequent use cases for GenAI in the business world is helping with email writing, I got curious about whether a similar pattern might show up in Google Trends for searches like “How to write an email.” ?? For the analysis, I used an interrupted time series method, which examines the effect of an intervention by comparing time series data before and after the intervention that happened at a known point in time, allowing us to assess any shifts in level or trend. ?? As the resulting charts show, there does seem to be a noticeable drop in searches on how to write emails following ChatGPT’s official release. In fact, the shift is so pronounced that it’s clear even with a quick eyeballing analysis. Sure, we can speculate about other factors to be involved - maybe we're sending fewer emails because we’re relying more on platforms like Slack or Teams - but as other available stats suggest, we can only wish that were the case ?? Either way, CausalPy turned out to be a super easy-to-use and user-friendly tool. I'm looking forward to using it in my future projects. Kudos to the entire CausalPy team! ?? To replicate the analysis, you can find the code on my blog here: https://lnkd.in/eauD3TV8 #causalinference #causalpy #genai #emailcommunication #python

I’m excited to continue my series on Marketing Mix Modeling (MMM) with my latest article being published by Towards Data Science, focusing on Model Calibration with PyMC! This is a crucial topic that many organisations struggle with, often hesitant to acknowledge that models based solely on observational data are biased. Understanding and addressing these biases is essential for making good marketing budget decisions. Let me know your thoughts! https://lnkd.in/eBE9fqrQ

??? In a recent chat, Ulf Aslak Lai reflected with Thomas Wiecki, PhD on moving from academia to industry. ?? In academia, Ulf thought, “This is just how people work—digging deep to fully understand problems.” But in industry, he found speed and surface solutions often took priority. It felt strange, even frustrating. ?? At PyMC Labs, he says it feels like “coming home.” Here, solving problems means truly understanding them, not just quick fixes. ?? If you’re looking to make your marketing strategies more effective, we’d love to connect and explore how we can support you. ? ?? https://lnkd.in/gqPWwbep #Interview #PyMCMarketing #CausalPy #Podcast

Christopher Fonnesbeck killing it on stage PyData NYC, teaching advanced #GaussianProcesses with PyMC ?? @ PyMC Labs

??? Join Alexandre Andorra , co-founder of PyMC Labs, and marine ecologist ?? Aaron MacNeil for a special live recording of the Learning Bayesian Statistics podcast at PyData NYC! ?? In this session, Alex will explore Aaron’s latest research on tracking the trade of shark meat between countries, diving into challenges like missing data, causal inference, and Bayesian methods. ??There will also be a live Q&A?and time to connect with others interested in Bayesian stats ??. ? Plus, stay tuned for another exciting LBS podcast event at PyData NYC with a soon-to-be-announced speaker! ?? ??? Event Info: ?? https://lnkd.in/gD3E5wD7 ?? https://lnkd.in/gR-8zc-4 #PyMCLabs #LearningBayesianStatistics #PyDataNYC #PyMCMarketing