Have you considered using a universal semantic layer in your analytics architecture?

We already have this12%

Yes, we are pursuing this53%

Yes, but we are not currently pursuing this14%


I’m not sure what that is4%


923 views3 Comments

Director of Data Architecture in Media, 5,001 - 10,000 employees
Modern data architectures require a semantic layer and oversight that ensures precision, reusability, reliability, and cost-effectiveness. The semantic layers should abstract and augment metrics, features, and other semantic components into semantic physical stores within API economic practices, allowing business users to consume and explore data without having to be technical data pipeline experts.

Symantec layer is also a governed, curated metrics/features data mart for business users, data analysts and data scientists to produce their analysis with more focus on the insights and less focus on backend tables structures. Semantic modelling enables collaboration toward measured objectives and doesn't impede it.
Director of Data in Healthcare and Biotech, 10,001+ employees
Yes, we currently employ a universal semantic layer for analytics. This is part of our initiative to "lower the skill floor," so to speak. I believe it's absolutely necessary to semantically redefine fields and structures, thereby creating a new baseline when there are dozens of heterogeneous data sources in play. Take into account the numerous ad-hoc reporting requests we receive for research and clinical operations. These requests must be manageable enough to be fulfilled within a day by a single individual, to maintain efficiency. The only feasible way to achieve this is by redefining the data structure.

In presentations, I often illustrate our process of redefinition with a classic example. Let's consider a scenario where someone submits an ad-hoc report from Epic, requiring cancer treatment regimens that begin with one drug and are followed by another, for all patients diagnosed with a specific disease within a certain timeframe. Moreover, they want the data broken down by race, ethnicity, and age. To accomplish this directly in Epic's Clarity would require navigating through numerous tables via complex joins. Even then, some of these tables bear obscure names that bear little relevance to their content, like the treatment table which is named something entirely different.

Now let's consider a new staff member tasked with this assignment. The learning curve for them would be steep. However, with our semantically redefined system, the tables are named intuitively (Demographics, Treatments, etc). This not only expedites the reporting process but also improves comprehension for newer staff members.

You can find a detailed explanation of this process in one of our publications: (https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000036).

In my experience across various groups and roles, this semantic layer approach is incredibly efficient. The solution doesn't necessarily lie in vendors or products; rather, the focus should be on establishing conventions, among other decisions.
Head of Data Strategy in Software, 51 - 200 employees
Semantic layers are increasingly a critical component of a modern data stack - but so too are MDM systems, which also act as a form of semantic layer.  Semantic layers link data at a schema level in support of analytical uses cases, while MDM's will link data at record or field level and support both analytical and operational uses.  For most larger companies with highly diverse data ecosystems, both are required.  

Content you might like

Insider threats – rogue admins19%

Encrypting my data51%

Deleting my backup copies11%

Resident malware8%

Data theft – data exfiltration11%



1.6k views1 Comment

Proactively seeking ways to integrate them13%

Evaluating their potential but proceeding cautiously80%

Not actively considering them in architecture7%

Not sure0%



CTO in Software, 201 - 500 employees
Without a doubt - Technical Debt! It's a ball and chain that creates an ever increasing drag on any organization, stifles innovation, and prevents transformation.
Read More Comments
40.7k views131 Upvotes319 Comments