Adrian Antico

We're facing more and more questions about the use of synthetic data in market research. I'd recommend this blog post, which really resonated with me. So much of the conversation is "how closely can this replicate human data?" But there are so many other considerations researchers should be thinking about. https://lnkd.in/eSQfSTPj

31

3 Comments

You trained the model. You trusted the output. But can you trace the data that led to the decision? 👀 AI in analytics isn’t dangerous because it’s wrong. It’s dangerous because it’s convincing, even when it’s wrong. 📰 This article breaks down the real risks of AI-powered insights and how to design for trust with traceability, governance, and human oversight. Read it here: https://lnkd.in/eKuQnv7J

10

The Databricks Startups team answers founders’ most common AI questions—where to start without hype, how to stay compliant with sensitive data, what to build versus buy, how to cut hallucinations, and how to keep costs in check. 💡 Get step‑by‑step starting points, governance patterns for sensitive data, a pragmatic build‑versus‑buy lens, tactics to curb hallucinations, and cost controls that scale. 🔗 Full post: https://lnkd.in/ehRa3Yp5 #Databricks #DatabricksForStartups #AI #GenAI #MLOps

63

I love this approach which I use often. Do not show off your technical prowess, on how fancy you can get, NO. Solve a real problem and facilitate a high impact decision making from the data/statistical science solution you offer.

1

Excited to share a little side project I've been tinkering with: statlingua! 🗣️ For those of us who spend a good chunk of our time translating statistical concepts into plain English (whether it's for stakeholders, collaborators, or even just our future selves!), I thought it would be neat to have a tool that could help automate some of that. statlingua is a pretty simple R package that aims to generate natural language descriptions of common statistical models and results. Think of it as a way to get a first draft of an explanation for things like regression coefficients, model summaries, and maybe even some basic visualizations down the line! It's still very much in development, so expect rough edges, potential inaccuracies, and a healthy dose of "coming soon" features. However, I'm putting it out there in the open in case anyone else finds the idea intriguing or wants to contribute. You can check out the (very basic!) code and contribute at https://lnkd.in/g3py6d6f. #rstats #datascience #statistics #opensource #genai #LLM

55

2 Comments

If your data/analytics team isn't leading the charge on AI, you're doing your organization a disservice. Tris J Burns 🔥 wrote about this in his newsletter and I wholeheartedly agree. Let’s be clear... it’s not because AI is going to replace data analysts (or anyone else). So why? Two really good reasons: 1. Someone, somewhere in your organization is using it, and they need your help. Data teams know the data, know the limitations and the nuances, how to be rigorous, and how rigorously we need to validate data before drawing conclusions . 2. Data teams know where the bodies are buried. All those weak data pipelines, poor metadata and workarounds you've been complaining about for years impact AI. No-one knows how to fix them better than you. I have many reservations about AI and the conclusions it draws, but used right, it's a fantastic tool. I use it, I've weaseled my way into several AI projects (mostly to annoy Peter G. McDermott), and regularly share my perspective on what we need to do to our data to get it working right. As a data person, you should be doing the same, otherwise your AI strategy will be built on weak foundations. What is your team doing to leverage AI or support AI strategy?

84

17 Comments

Stop analyzing clean data. It’s hurting your career. 🛑 Most data portfolios look the same: ❌ Perfect, clean CSV files. ❌ Generic topics (Titanic, Iris, Housing Prices). ❌ Zero business context. ❌ "Here is a chart because I can make a chart." Base Analytic was built to break this cycle. We don't just host datasets; we simulate the actual job. Here is how we are different: ✅ The Data is Messy: Real-world data is never perfect. Our datasets have missing values, timezone conflicts, and duplicate rows—just like production databases. ✅ Context First, Code Second: You aren't just writing SQL. You are answering to a CTO who needs to know why revenue dropped 15% in Brazil. ✅ Role-Playing: You step into a specific role (e.g., Lead Data Consultant) with specific clearance levels and stakeholders to satisfy. ✅ Full-Stack Workflow: Combine SQL for extraction and Python for analysis in a single, cohesive mission. If you are tired of "tutorial hell" and want to build a portfolio that actually impresses hiring managers, this is for you. Start your first mission for free: 👇 [Link in comments] baseanalytic.com #DataAnalytics #BaseAnalytic #SQL #Python #PortfolioProject #DataScience #CareerGrowth

4

2 Comments

Building agentic workflows? There are actually only a few key concepts to be aware of. StateGraph orchestration: A StateGraph is the agent’s blueprint. It consists of nodes that handle tasks with edges (directions) to other tasks, defining the order. Together, they provide the agent with a flow to follow. Conditional routing: At decision points, the agent inspects results and follows “if-else” statements that skip failures, handle edge cases, etc. Memory and threads: Thread memory holds recent conversation turns. Essentially, it stores the short-term memory that you are John Doe and were interested in the weather. Long-term memory stores user facts, letting the agent stay context-aware, reduce repetition, and personalize replies, such as that you have been a customer since 2018. Human-in-the-loop: Strategic hand-offs let people review low-confidence outputs, catch critical paths or errors, and feed corrections back to the model, keeping quality and alignment tight. If this still feels too abstract, I’ve included a more detailed explanation in the article below. https://lnkd.in/eAr-6YAQ

14

1 Comment