Production MMM breaks when influencers aren't interchangeable, campaigns shift, and channels enter mid-series. Part III of our Nürnberger case study tackles all three. 27%+ CPL reduction: dub.sh/5LLFAJ9 #MarketingMixModeling #BayesianStatistics #PyMC

Last call! Our Master Agentic Data Science course starts tomorrow. Build agent-assisted EDA, causal inference pipelines, and decision systems. Code ADS10L for 10% off. 👉 dub.link/9YG77NC #AgenticAI #DataScience #PyMC

We're sponsoring the M-Agents Hackathon at #NYCTechWeek on June 7. Go from zero to a working multi-agent AI system in one day, built with open-source tools like #PyMC and Decision Lab. Winners get a seat in our Agentic AI course. Spots are limited dub.link/ZE70nxd

@wegmansofficial.bsky.social needed to quantify cannibalization risk before opening new stores. We built a Bayesian system that answers in minutes, not weeks. Full technical walkthrough: dub.sh/1ENBxEc #RetailAnalytics #BayesianStatistics #PyMC

The radar plot is the most popular chart in football analytics. It might also be the least effective. Chris Fonnesbeck breaks down why and builds the replacement: dub.sh/BUsCcev #SportsAnalytics #DataViz #Bayesian

Data science ≠ software engineering. Automate the wrong choices, and you just get confident wrong answers faster. Tomorrow: @twiecki.bsky.social + @hugobowne.bsky.social dig into what agentic data science actually looks like. 👉Register here: luma.com/juquutuy #AgenticAI #DataScience

We mapped the AI-course landscape for data scientists. One area was underserved: Bayesian agents for decision-making, where statistical rigor matters as much as speed. That's why we built the "Master Agentic Data Science" Course dub.link/DVJD47D #AgenticAI #DataScience

We replaced a PyMC compiler with an LLM agent that transpiles to Rust. 3-7x faster sampling, validated against reference gradients, ~500 lines of scaffolding. 👉 Full technical dive: dub.sh/4QNM59l 🧩 Open source: github.com/pymc-labs/alchemize #PyMC #Rust #AgenticAI

"What if I shift 20% of TV into paid social?" Tomorrow we're typing that into a live MMM agent. Real answer, under 30 seconds. Join us live: dub.sh/EbAc1NH #MMM #Agentic

We're offering free scholarship seats for our #Agentic Data Science course. Grad students, early-career data scientists, nonprofit researchers, practitioners where course fees are a barrier -- we don't want cost to stop you. Apply before May 27: ggl.link/EubkG6l

"How many people did COVID kill in the UK?" is the wrong question. The right one is counterfactual: how many would have died without it? You can't observe that world. You have to model it. #BayesianStatistics #PyMC #CausalInference

I want to use AI agents, but I can't trust the output." We kept hearing this, so we built a community to solve it together. Daily AI news briefs, early tool access, and real conversations. The Decision AI Discord is open: discord.gg/mfNmfhhwhR #DecisionAI #AgenticAI #PyMC

"What happens if I shift 20% of TV into paid social?" On May 7, we're typing that into a live MMM agent. It answers in under 30 seconds. Register to join us live: dub.sh/HJ7h4iu #MMM #AgenticAI

Why do women outlive men by 12 years in some countries but only 2 in others? @allendowney.bsky.social breaks it down with Bayesian counterfactual modeling, live at @odsc.bsky.social East, April 30. #ODSC2026 #PyMC #BayesianStats

Why does LASSO shrink every sector's coefficient toward zero, whether momentum, value, or profitability? Wrong target. Partial pooling in #PyMC shrinks to a learned global mean instead. Net Sharpe 1.47 on US equities after costs Full blog here: dub.sh/icrbIgV #QuantFinance

5 years of UK electricity demand, decomposed in 60 seconds on a laptop. Trend + annual cycle + weekday effect, one compositional Bayesian model in PyMC with HSGP. That's what we're teaching in London, Jun 8–10. Early-bird closes Apr 27: dub.sh/GAjQB5gg #PyMC #Bayesian

Your new SKU's sales chart is up. Half the volume might be coming from your own flagship. A Bayesian discrete choice model traces every unit back to its real source: own-portfolio vs. competitor vs. brand-new buyer. Here's what that picture looks like ↓ #MMM #CPG

Most MMMs sit idle because only three people in the org can drive them. The fix isn't a better model. It's a better interface. Agentic dashboards, live demo: May 7, 11 AM EST. Signup here: dub.sh/2VOLNC8 #MMM #MarketingAnalytics #AgenticAI

Nürnberger had a classic #MMM problem: lower-funnel capped, upper-funnel invisible, GDPR shortening journeys. We built a funnel-aware MMM in pymc-marketing with censored demand + dual likelihoods. Result: 27%+ lower CPL in 6 months. Read Part I: dub.sh/o690ugv #Bayesian

An uncalibrated #MMM can get ROAS direction wrong, not just magnitude. We walk through prior predictive checks, ROAS parametrization, saturation likelihoods, and where lift tests break down. Full recording available on youtube : dub.sh/oAjXCiG #MMMCalibration #BayesianModeling

Featured case study from the community: a two-pillar Bayesian CLV framework for a nonprofit, built on pymc-marketing. 1M+ records, MCMC vs. MAP tradeoffs, and a full Databricks deployment walkthrough Read the blog here: dub.link/AO4oazi #PyMC #BayesianStatistics #CLV

Watched your MCMC sampler fail after 20 min? INFERLOG HOLMES flags convergence issues in real time, 30 seconds into sampling. Full webinar recording with the researchers + Chris Fonnesbeck and Oriol Abril Pla now available to watch on YouTube: 👉 dub.link/v7W5hwI #BayesianInference

Getting an LLM to run an analysis is easy. Getting it to run one you'd stake a decision on? Different problem entirely. 4 live sessions. Spec-driven workflows. Bayesian rigor. June 2-11. Signup here: dub.link/kcIhjOi #AgenticAI #DataScience #PyMC

Your model was right. But it was too slow to matter. AI agents can sit on top of your Bayesian models and give stakeholders answers in seconds, not days. New from #PyMCLabs on the decision bottleneck: dub.link/odtRwW7 #DataScience #AIAgents #BayesianModeling #AgenticAI

Equity markets cycle through regimes (low-vol trends, selloffs, grinds), each with different return profiles. Averaging across them fits a market that never existed. Bayesian HMM in PyMC. 94% regime detection accuracy. full dive: dub.link/uRxiSsY #PyMC #QuantFinance

Most data scientists learn Bayesian stats from textbooks. Few get to learn from the people who built the library. This June, @fonnesbeck.bsky.social Niall Oulton & Robert Zinkov are running a 2.5-day hands-on PyMC workshop in London. Reserve your spot: dub.link/oaudJAt

Your AI agent recommended shifting 30% of your budget to TikTok. The code ran, the model converged, and the viz looked clean. The recommendation was wrong. We built decision-lab to fix that, and we're open-sourcing it: dub.link/4wNiUdX #AgenticAI #OpenSource #DataScience

Your optimization model is only as good as the forecast you feed it. We partnered with FICO to show what happens when you replace point estimates with full distributions. The results changed the decisions, not just the forecasts. Full technical dive here: dub.link/zrDyPCF #PyMC #FICO

Your #MMM says paid search delivers 4x ROAS. Your lift test says 1.8x Which drives your budget? Stop choosing. Integrate lift tests as likelihoods to calibrate MMM. More defensible ROAS. Real uncertainty. Better decisions. Register → dub.link/aehzbmX

We're bringing Bayesian modeling to London, live and in person! June 8-10: Learn to build probabilistic models in #PyMC, guided by its creators. Small cohort, hands-on, walk out with working code. Seats are limited. 👉 Reserve your seat: dub.link/86tcbuP #BayesianStatistics

Tired of AI confidently making bad analytical decisions? Meet Decision Lab, our new open-source harness for #agentic data science. It explores multiple analysis paths, uses #Bayesian diagnostics as feedback, and stops when assumptions fail. 🧩Github: dub.link/wKpAWNi

The Applied Agentic AI Data Science course is here – 12 hours of live, hands-on training to build reproducible AI workflows. Learn directly from Thomas Wiecki, Hugo Bowne-Anderson, and Luca Fiaschi. Secure your spot → dub.link/xriAsX6

An MMM can predict perfectly, and still get causality wrong. This chart shows it: uncalibrated ROAS posteriors miss reality. Add one lift test → estimates snap back to truth. Join our MMM Calibration deep dive Webinar (Apr 8): dub.link/Nax7e44

Most #MMMs explain the past but choke on the future; they need variables that haven't happened yet. Our latest blog by Dr Luca Fiaschi shows how pairing PyMC-Marketing with Chronos2 fixes this: causal clarity + forecasted controls, just +0.34% added error. 👉 dub.link/jsnWOzR #PyMC

Is your MMM anchored in reality, or just an approximation? On April 8, four PyMC Labs data scientists share the calibration playbook: lift tests, prior modeling, geographic validation & more. 👉 Register here: dub.link/dB5mPlA #MMM #MarketingAnalytics

New research proposes debugging #Bayesian inference while it runs, not after. A methodology for live MCMC diagnostics: R-hat, ESS, and trace plots updating in real time instead of post-hoc. Join our webinar on Mar 25, 11 AM ET. 👉 dub.link/Bwr7kT3 #PyMC #MCMC #DataScience

Most quant models are correlational - they tell you what moved together in the past. But robust investing needs more than correlation. It needs causal structure + functional form. Our latest blog explores how the two work together. 👉 Read more: dub.link/Xe9cHWg #CausalInference #QuantFinance

LLMs give surface-level MMM advice So we built AI skills for pymc-marketing that know how to set priors, handle adstock carryover, and run budget optimization. They ship with the repo and work in #Cursor/#Claude 👉Github Repo: dub.link/e5qeoQE #PyMC #MediaMixModeling

Our Sports Analytics Team is heading to #SSAC26! Join us for a hands-on workshop on Bayesian Spatial Modeling of #NHL goaltending. From shot difficulty to rink-wide scoring probability, see #PyMC in action! 👇 dub.link/qMEf0Ni #SportsAnalytics #BayesianModeling #Hockey

We gave Claude a 15-page PyMC best-practices cheat sheet. Pass rates jumped from 60% → 93%. Hardest tasks saw the biggest gains: • Stochastic volatility: 0% → 67% • Runtime: 19 min → <3 min and more .. 👉Benchmark: dub.link/KZxxZfQ 👉 Get the skill: hub.decision.ai/skills/pymc-... #BayesianModeling

Your next focus group might be synthetic🤖 Traditional Research is slow, expensive, and limits testing. Synthetic consumers help you simulate feedback and iterate faster. We put together a practical guide on #SyntheticConsumers : dub.link/jczwvFL #MarketResearch #AI

We’ve been building agentic data science systems; the hardest part isn’t execution, it’s knowing if an agent’s conclusions can be trusted. Join our webinar on Feb 25 with #PyMC Labs to learn how we catch failures & build trust 👉 dub.link/SUHEjct #AgenticDataScience #MMM

How do #MLB teams evaluate pitchers behind the scenes? ⚾ We use Bayesian Additive Regression Trees to isolate pitcher skill from defense and park effects. See why BART outperforms XGBoost for building pitch quality metrics 👉 dub.link/McvDt9S #Baseball #SportsAnalytics

#Bayesian causal inference is about generative mechanisms, not just correlations. Learn how to encode causal assumptions and simulate counterfactuals in our Applied Bayesian Regression Modeling course. Registration still open 👇 dub.link/PKrUYFP #PyMC #CausalInference

Bayesian regression shouldn’t stop at good fit and nice plots. Our Applied Bayesian Regression Modeling course shows how to build models that drive real decisions using #PyMC & #Bambi 👉 Learn more & enroll: dub.link/x1K1vz4 #AppliedRegression #Analytics #DecisionMaking

PyMC-Marketing just hit 1M downloads on GitHub 🙌 Huge thanks to the open-source community; this is just the beginning. Explore the repo 👉 dub.link/KB2XwIK #PyMCMarketing #OpenSource #MarketingScience #MMM

Stop testing AI agents like software. Start testing them like Data Scientists. Learn how we build scientific guardrails for agent reliability. Register for our Feb 25 webinar 👇 dub.link/p9IiuT1 #GenAI #AgenticAI #AIEngineering #BayesianModeling #PyMC #MLOps

Stop burning budget on ad creatives that haven’t been tested. A/B test your ads against synthetic U.S. consumers before you launch! Less guesswork. Lower risk. 👉 Test your creatives: dub.link/oFT1mle #AdTesting #MarketingScience #SyntheticConsumers

Learn Bayesian regression modeling from the people building #PyMC, #Bambi & #CausalPy. 4 weeks. Live sessions. Real-world use cases. 👉 Registration now open: dub.link/2CDdYLK #BayesianStatistics #Baysian

Bayesian #MarketingAnalytics isn’t about complicated math. It’s about making decisions under uncertainty. Our #PyMC Labs’ 5-step MMM workflow shows how to go from raw data → probabilistic decisions → real-world action. 👉 Join the course: dub.link/zxxlnd6 #MMM #Bayesian