Longhow Lam

You have your KPI metadata sorted like a pro to scale your Tableau #KPI analytics? Then this is for you: #PowerKPIs Metadata Formatting ❤️ Excited to share a new feature Merlijn Buit and I have been working on this week and that will be available soon. It's easy like that: 1️⃣ Add your metadata (e. g. number formats, decimals, chart types) to the view 2️⃣ Map your metadata fields to the KPI attributes 3️⃣ Save and enjoy! Infotopics | Apps for Tableau Richard Van Wijk Siebe Grijpma

579

25 Comments

This week, our auditors will delve into the diverse world of data products, gaining insights into how different types are used to drive value in organizations. Understanding the nuances between these products is key to mastering data product management.

1

🚀 Unlock the Full Potential of Your AI with RAG! 🚀 🔍 Retrieval-Augmented Generation (RAG) is the game-changing technique that blends Large Language Models (LLMs) with real-time knowledge retrieval to deliver smarter, more accurate results. Ready to supercharge your AI workflows? Here’s your go-to guide for RAG Workflow Optimization Tips: 🔑 Key Insights to Master RAG: 1️⃣ Evaluation: Ensure your AI meets your performance goals. Test for general accuracy, domain-specific needs, and retrieval capabilities. Don’t leave success to chance! 2️⃣ Fine-Tuning Magic: Discover the best fine-tuning strategy for your LLM—be it Disturb, Random, or Normal initialization. Custom optimization = superior results. 3️⃣ Summarization Simplified: Go Extractive with tools like BM25 or Contriever for precise summaries. Opt for Abstractive methods (e.g., LongLLMlingua) for creative and context-rich outputs. 4️⃣ Smart Query Classification: Equip your model to classify queries dynamically, unlocking tailored retrieval strategies for every use case. 5️⃣ Cutting-Edge Retrieval Techniques: Blend BM25 for keywords and Hybrid Search (HyDE) for embeddings. Simplify complexity with Query Rewriting or Query Decomposition for nuanced queries. 6️⃣ Embeddings That Work: Leverage powerhouses like intfloat/e5, Jina-embeddings-v2, or all-mpnet-base-v2 for top-tier vector representations. 7️⃣ Vector Databases: Store and retrieve embeddings effortlessly with robust tools like Milvus, Faiss, or Weaviate. 8️⃣ Repacking & Reranking: Transform retrieval quality through repacking and reranking with state-of-the-art models like monoT5 or RankLlmAM. 🌟 Why RAG is Your Next Big Leap: RAG doesn’t just generate—it evolves. By integrating external knowledge in real-time, it unlocks potential across tasks like: ✅ Advanced Question Answering ✅ Domain-Specific Solutions ✅ Dynamic Document Summarization 💡 Pro Tip for Success: Master the trifecta of chunking, embedding selection, and retrieval strategies to scale up your RAG pipeline with ease. 🔥 Don’t Miss Out! If you’re not integrating RAG, you’re leaving potential untapped. This is your chance to deliver smarter, faster, and more accurate AI solutions. 💬 What’s Your RAG Strategy? Share your challenges and wins with RAG in the comments. Let’s collaborate and create cutting-edge solutions together! #AIInnovation #RAGOptimization #AIWorkflows #EmbeddingTech #AIforGood

6

PowerKPIs allows all your users to access your KPI grids and driver trees throughout the entire organization on every level. With this enhanced meta data embedding you lift your steering information to the next level. Want to know more? Contact me!

3

If you think KPI cards in #tableau are great but require a lot of timw/effort to build, have a look at POWERKPIs It is basically a drag and drop KPI card and driver tree builder.

3

I've been seeing many 'essential skills' lists about working with data on LinkedIn. These usually focus on technical skills and tools. Technical skills are crucial for data professionals. They often neglect a crucial piece: 💡 understanding the business context and how data can drive real value As data experts, we can unlock powerful insights that shape strategic decisions and drive organisations forward. Our true impact comes from bridging the gap between data and tangible business outcomes. 💡 Technical skills alone are not enough. We must understand our company's goals, challenges, and data's role in addressing them. This may also mean understanding our Customers goals and challenges too. By understanding the business context, can spend our time on actionable intelligence. Data pipelines and analytics that resonate across the organisation. This deeper understanding helps us: ✔ ask the right questions ✔ identify critical patterns ✔ deliver solutions aligned with company objectives When we grasp the bigger picture, we can anticipate future needs and drive innovation. We become strategic partners, shaping our organisations' direction and unlocking new growth and competitive advantage opportunities. To elevate our value as data professionals, we must embrace a mindset beyond technical expertise. Let us actively seek to understand our business, customers, and industry landscape. #DataLeadership #StrategicInsights #DataAnalytics #CompetitiveAdvantage

20

1 Comment

🎈𝟰𝟭 𝘁𝗼𝗱𝗮𝘆 - and of course I asked ChatGPT for a birthday roast based on my previous chats... 👇 𝘾𝙝𝙖𝙩𝙂𝙋𝙏: "At 41, you're entering that phase where hype isn't just something you critique in your writing, it's also the sound your knees make when you stand up too fast. I imagine your technology pamphlet will now include a chapter on how to avoid the hype of new joints, followed by one on the pitfalls of chasing "innovations" like a good night's sleep. They say life begins at 40, but given your philosophical leanings, you're probably still in the middle of a deep reflection on whether life ever really started, and if the term "begin" is even valid. In any case, you’re at the age where wisdom meets… a reminder to schedule that annual checkup. In your podcast, The Only Constant, you often muse on change, but ironically, your caffeine addiction and Arsenal loyalty remain steadfast. Even though Arsenal is probably more committed to giving you heartaches than victories, we admire your loyalty to both. Maybe by the time you're 50, they'll let you enjoy a title! So here’s to 41—where hype cycles are shorter, but the hangovers last longer. Keep being the sharp, philosophical business guru who’ll continue roasting the latest trends… because someone’s got to keep these young execs grounded while simultaneously teaching them how to survive with chronic tech whiplash. Happy birthday, mate! Cheers to many more years of brilliance, and may you continue naming everything (except, perhaps, those gray hairs sneaking in)!" . . . . And I actually asked it to make it kind... 😅 #birthday #roast #chatgpt

93

53 Comments

I used to think complex models were everything in AI. Then I learned: real value starts with simple, goal-aligned data. Simplicity drives results. 🧩 Want to hear more lessons from my data journey? Hit follow! #AIstrategy #DataDriven

5

1 Comment

If only that was a solveable problem... Well... the problem is not only to manage the data itself, but a culture and process problem even more.

4

Missed our livestream about Dynamic Queries in M? Im happy to say I presented at the Persian usergroup with Vahid Doustimajd. In the session we had a thorough walkthrough about areas that can break your queries. You can now see the recorded version at: https://lnkd.in/e2GVegft If you missed it, you can still look back. Question for you: What other areas do you look at to keep your queries ?

91

9 Comments

In today’s blog post, I repost an ungated version of my Business Day article with Pietman Roos in which we discuss the importance of data ethics in the age of big data and AI. https://lnkd.in/dTN86aiR

5

1 Comment

🚨 Streaming and Push semantic models in #PowerBI are being deprecated. Anshul Sharma and I presented at European Microsoft Fabric Community Conference on RTI for Power BI developers where we talked about how the RTI in Fabric provides scalable and performant options. You will not be able to create any new streaming semantic models in 2 weeks which is a bit too aggressive IMO but if you haven't looked at RTI in Fabric yet, now is the time ! https://lnkd.in/g4zBNswb

50

3 Comments

Moving a 2000 workbook environment with multiple data sources to #TableauCloud with TabMove. We claim to be 10x faster than the manual migration process. Don't take our word for it: "The actual migration took place on a Friday, was completed end-of-day, enabling the full user base to switch to the new Tableau cloud sites." Read more here about how TVH moved to cloud seamlessly: https://lnkd.in/eKZpX45v

59

1 Comment

Do you know that feeling when you talk to people, and it feels really comfortable and good? I experienced that while participating in the Betabit podcast, where we discussed how to bridge the gap between Dataviz and Business Intelligence. Rick van den Bosch and Oscar van Tol are such great people, and I’m really happy they invited me to participate! The conversation went beyond data visualization and BI to address challenges like navigating tech as a woman and turning technical knowledge into real-world impact. Reflecting on creativity, community, and balance was a great reminder of what drives meaningful change in our industry. Also, huge thanks to Mies Koopman, for the smooth coordination and for making sure I felt at ease before we started recording. 🎥 𝐖𝐚𝐭𝐜𝐡 𝐭𝐡𝐞 𝐯𝐢𝐝𝐞𝐨 𝐩𝐨𝐝𝐜𝐚𝐬𝐭 𝐡𝐞𝐫𝐞: https://lnkd.in/ejDemRe6 🎧 𝐋𝐢𝐬𝐭𝐞𝐧 𝐭𝐨 𝐭𝐡𝐞 𝐩𝐨𝐝𝐜𝐚𝐬𝐭 𝐡𝐞𝐫𝐞: https://lnkd.in/eiirVwRF #DataVisualization #BusinessIntelligence #TransformingData #PowerBI

33

2 Comments

🚀 DE Interview📢Data Imputation⌛1-Min-Read 🧔‍♂️Interviewer: Can you explain how you improved data accuracy in your project, leading to significant business impact? 👶Candidate: In our project, we were facing issues with inaccurate business results due to missing values in our raw data layer. We knew which records were missing values and had a clear understanding of what those values should be. To tackle this, I developed a data imputation framework and created a new layer as a parallel pipeline. Here’s how I approached it: Data Imputation: I designed a framework that identified missing values and filled them with the appropriate data. This involved analyzing patterns and ensuring the imputed values were accurate and relevant. Parallel Pipeline: I set up a parallel pipeline to compare the imputed data with the actual data. This allowed us to validate the effectiveness of the imputation process. Comparison and Validation: By running the imputed data through the pipeline, we could see that the imputed data was much closer to the actual results compared to the raw data. Production Implementation: Due to the significant improvement in data accuracy, the team decided to move my imputation code into production. This ensured that business results were more accurate and reliable. By implementing this data imputation framework, we achieved several impacts: Increased Accuracy: The imputed data significantly reduced discrepancies, resulting in more accurate business insights. Improved Efficiency: Automating the imputation process saved time and resources, allowing the team to focus on more critical tasks. Enhanced Reliability: The production implementation of the imputation framework ensured consistent and reliable data quality. Real-time Validation: The parallel pipeline provided real-time insights, helping us quickly identify and address any data issues. Summary: 1️⃣ Problem: Inaccurate business results due to missing values in raw data. 2️⃣ Solution: Developed a data imputation framework and parallel validation pipeline. 3️⃣ Impact: Increased accuracy, improved efficiency, and enhanced data reliability. #DataEngineering #DataImputation #DataAccuracy #Automation #Python #DataValidation #TechInnovation #BusinessIntelligence #DataQuality #Efficiency

22

1 Comment

🚀 I really want to tell you about the cool stuff we’ve been making in Future:Lab - and what we’ve learned. Future:Lab is Oakland | B Corp 's R&D department 🔬, and it’s our job to figure out and prototype what we think is coming next in the data space 📊. Obviously, right now that’s generative AI 🤖 - but we’ve taken it one step further.We’ve built out an intelligent agent framework which arms large language models with the planning space, knowledge, and tools to actually take actions and solve problems in the real world 🌍. Seriously though, I can be quite boring about this - so please message me if you’d like a chat about the exciting stuff we’ve been building 💬. Today though, I want to tell you about one of the things we’ve learned in working extensively with LLMs over the last two months 🧠. The most important insight we have had is that the instructions and tooling provided to AI models should be as simple as you can make them - without loss of function 🛠️. However, simplicity doesn’t mean lack of sophistication. It means creating an intuitive interface that the AI can understand and execute without ambiguity 🤓. Clear instructions and well-designed tools lead to more reliable and efficient outcomes, much like how comprehensive training helps human employees excel in their roles 🏆. We have found the same problems typical of humans set tasks in teaching our synthetic counterparts to do their jobs. A really great example of this is our agents that use code generation to solve their problems 💻. We have a suite of data analysis agents (we call them Dave!), which can write and execute Python to draw insights from the data provided to them. We have found that when they are faced with a blank slate, they can run off into the distance and write code not necessarily relevant to the problem. They then have overseer agents which check their work and send it back if it’s rubbish 🗑️ - but why waste all that time? If we tell Dave which code libraries and strategies to use ahead of time, we find we get much better responses! 🎯 If this sounds interesting to you, then we’re industry-leading here in Future at Oakland in developing AI Agent capability for business 💼. ❗️ Want more insights like this? Check out our free, Ultimate Guide to AI. Hit the link in the comments! 🔔 Want more of this content in your feed? Don’t forget to hit the bell on my profile.

25

2 Comments

PLEASE reshare on your #LinkedIn timeline if you haven't already- it means THE WORLD to me- thank you! #AI/ #DataScience/ #MachineLearning: FREE Infographic, The Neural Network Zoo by the Asimov Institute and Fjodor van Veen!!! From the original article: "With new neural network architectures popping up every now and then, it’s hard to keep track of them all. Knowing all the abbreviations being thrown around (DCIGN, BiLSTM, DCGAN, anyone?) can be a bit overwhelming at first. So I decided to compose a cheat sheet containing many of those architectures. Most of these are neural networks, some are completely different beasts. Though all of these architectures are presented as novel and unique, when I drew the node structures… their underlying relations started to make more sense." #neuralnetworks #analytics #artificialintelligence #bi #bigdata #bigdataanalytics #data #dataengineering #datainsights #datascientist #datamining #dataanalytics #deeplearning #digitaltransformation #innovation #management #ml #technology

202

45 Comments