Guido Schauwecker

GenAI requires quality data to deliver, but what do you do if your data quality is poor. Here are a few pointers ... ;-) https://lnkd.in/dhhF5jPn

1

When it comes to data platform strategies, the expectations between executives and data leaders can sometimes feel worlds apart. CEOs might want the latest buzzword-heavy architecture, like "centralized data mesh fabric," without fully understanding what that actually means for implementation. Meanwhile, the head of data is left wondering how to explain that these architectures aren't just a one-size-fits-all solution. It’s all about aligning vision with practicality building a strategy that not only sounds impressive but actually works for the business's unique needs.

6

1 comentario

🔍 Roche’s Maxim of Data Transformation: "Data should be transformed as far upstream as possible, and as far downstream as necessary." I recently revisited an insightful article by Matthew Roche on a principle that resonates deeply with data professionals: Roche's Maxim. This guideline offers a strategic approach to data transformation, emphasizing the importance of balancing efficiency and cost in data processes. You can read the full article here: https://lnkd.in/gdr6F7RC 💡 What does it mean? Upstream refers to transforming data closer to where it’s generated. This approach often reduces costs, since transformations happen less frequently, and the value added can be leveraged across multiple use cases. Downstream means transformations closer to where data is consumed, often necessary when the transformation is context-specific, like user interactions or filtering in reports. Why is this relevant? In practice, it helps answer common questions like: Should you transform data in Power Query or DAX? 🛠️ Start with Power Query, as it’s further upstream. Should you transform data in Power BI or the data warehouse? 🏢 Opt for the data warehouse, ensuring all downstream systems work with a consistent data version. By transforming data at the right stage, we not only improve performance but also ensure a consistent data experience across analytics solutions. This principle can apply across different tools and platforms, from Azure Data Factory to Power BI. 🚀 As I continue to work with clients on building efficient data solutions, I see the importance of applying Roche's Maxim to create more effective and scalable data pipelines. #DataTransformation #BusinessIntelligence #DataAnalytics #PowerBI #AzureDataFactory #DataModeling #DataGovernance #RochesMaxim

7

Excellent and concrete article that explains by example, the combination of RAG and Graph database to extract information from heterogenous sources (structured and not) that are linked together directly or indirectly, with performance increase by vector-based indexes: #neo4j #graph #vectorindex #langchain #embeddings #llm https://lnkd.in/eBy7_MZf

3

1 comentario