A whiteboarding session that describes the five stages (ingest, store, transform, model, visualize) that make up a modern data warehouse (MDW) and the Azure products that you can use for each stage.
Interesting concept there at the end: prototype in PBI, then get back to me. Would love to see a blog/video on advice for that you've done in the past. Great work, as usual, James!
what most people are not doing, is they just explain the concept and they are not explaining what tools to be used, where you did in your presentation. Thank you for this clear explanation. I want to implement this scenario and I am looking for sample data where I can also create a data model in target SQL DB. Do you know such a dataset to be used?
Thanks James. I was wondering how would you see a datavault approach in the DWH layer, considering different rules would be applied before creating the dimensional model.
Thanks James for this video. Explains things well. I wanted to ask where you would add on a real-time API capability to support a secure API marketplace for retrieving read-only data with sub-second response times. The use case would be a replica of a healthcare EMR.
Thanks for your insightful videos, James. Do you recommend getting the individual Azure products or is Synapse Analytics a better way to go now? Seems like Synapse would allow you to see everything in one place, while the individual tools you'd need to piece together everything more manually. Do I understand that right? Thanks again
You have it correct. Think of Synapse as a single pain-of-glass where all the tools live under one roof and you never have to leave Azure Synapse Studio
Great post. Been an avid reader of your blog series for years. Q: Re: Guidance for Synapse for Large Enterprises w/ mature BI: Are we now 'officially' past the era where we need to consider the ole 'hub-n-spoke' model for DW & Marts w/ separate Db-gear & architectures (hub-n-spoke = whereby you needed a DW for storage & adhoc/deep analytics, but separate SQL Srvs for domain/subject based star/models (with all their SCD characteristics & low-lat/hi-concurrency usage). What's the current guidance from MS on this?
Hi Eric...I can't comment on Microsoft's guidance but my thoughts are the traditional hub-n-spoke with the spoke being a separate database is rarely used anymore. Rather, the "spoke" is the tabular model (dataset) in a star schema in Power BI
Thanks James. Do architectures like this only work for large companies or do you think small-medium size businesses could benefit from this over something like a traditional on-prem model (staging-ODS-DW)?
Hi James, do you have some thoughts around MDM as it relates to improving the end results of a data warehouse and the overall data within the application architecture?
Hi Pete, MDM is a must in most data warehouse solutions. I always make sure to tell customers to put a MDM product in your project plan because you will need it
