The Data Modeling Layer
The data modeling layer — While writing The Analytics Setup Guidebook, we realized that the tools we recommend (apart from Holistics!) were really drawn from a small handful of ideas. For lack of a better name, we decided to call these tools 'data modeling layers', or 'data modeling layer tools'. This post explains what the identifying markers are, and how to spot them in the wild.
This should be useful to you if you've not kept up-to-date with the various approaches to data analytics stacks. We think more and more tools will adopt these ideas, and that the approach will be a dominant one going forward.
Links From Elsewhere
AWS launches Amazon AppFlow — We're a little late to the party on this, but AWS launched a new SaaS integration service on the 23rd of April, earlier this month. The idea is that you can transfer data between AWS and external SaaS applications like Google Analytic, Marketo, Salesforce, Slack, Snowflake, and Zendesk. This is going to greatly reduce the costs and complexity of loading data into an AWS-hosted data warehouse. I'm interested to see where this goes.
Building an Analytics Stack in 2020 — One of the major ideas behind our writing The Analytics Setup Guidebook is the observation that much knowledge on contemporary data stacks are trapped within the heads of practitioners. This post by Nelson Auner calls what we call a 'data modeling layer' a 'business logic layer', which goes to show that there's no good name for this yet. But we agree with pretty much all Auner says, though we go into way more detail in our upcoming book.
And so, about that book. Here's the cover, which our designer whipped up early on in our editing process:
I'm really quite excited about this. In the book, we go through the basics of setting up a minimal viable analytics stack, and then walk through the choices that we think best takes advantage of the latest techniques and latest technologies. Of particular interest to those who are experienced data professionals, we include a chapter that takes a first stab in blending Kimball-style dimensional modeling with more contemporary techniques.
Anyway, if you're interested in the book, hit reply and let me know. I'll send you an early copy.
And — even if you're not interested in the book — what do you think of this week's email? Do send me an email and I'll promise to answer.
I hope you're reading this safe, sane, and healthy.
Till next Wednesday,
Warmly,
Huy