Dimensional Modelling in the Modern Data Warehouse | Data Days 2022

TL;DR

Analyzing the transition from star schema to one big table, questioning traditional data modeling practices in modern data browsing.

Transcript

sorry this was part of the template um okay so this is uh the title of this talk uh i'm not gonna read it you can read it yourself um so there's this idea in my mind which is a question and i really don't have an answer for this question so i'm sorry if you were here for answers because so landscape technologically speaking in the area of data brow... Read More

Key Insights

• 😃 The shift from star schema to one big table prompts a reevaluation of traditional data modeling practices.
• 🇨🇷 Cost reductions in storage and advancements in columnar data browsing impact data modeling decisions.
• 🤩 Performance enhancements and improved understandability are key benefits of alternative data modeling approaches.
• 🔒 Security, maintenance, and hidden costs must be considered in adopting denormalized data models.
• 😊 Balancing the pros and cons of traditional vs. modern data modeling practices is essential for effective decision-making.
• ♿ Democratizing data access and simplifying data models can foster collaboration and innovation within organizations.
• 🪡 Adapting data modeling strategies to meet evolving technological needs is crucial for sustainability and efficiency.

Questions & Answers

Q: What inspired Mateo to question traditional data modeling practices?

Mateo was inspired to question data modeling practices due to changes in data technology and the prevalence of traditional methods like star schema.

Q: What are some reasons for using dimensional modeling like the star schema?

Dimensional modeling offers benefits like simplifying data models for human interaction, cost-effectiveness, and efficient analytical processing.

Q: How do costs factor into the comparison between star schema and denormalized models?

Mateo highlights that costs have significantly decreased with advancements in cloud computing and columnar data storage, impacting data modeling choices.

Q: What are the challenges and considerations when moving from a star schema to a denormalized model like the one big table?

Transitioning to one big table may pose issues like security concerns, performance optimization, and maintenance costs, requiring a thorough reassessment of data modeling strategies.

Summary & Key Takeaways

• Traditional data modeling practices, like the star schema, have been prevalent despite changes in data technology.

• Mateo, a data engineer, questions the efficacy of star schema vs. denormalized models like the one big table.

• Factors like cost, performance, and understandability play a role in rethinking data modeling approaches.