All the data and analytics insights for data leaders we've scooped up this week - 8/26/22.
Data Leader Digest - 9/30/2022
All the insights for analytics leaders we've scooped up this week. AWS vs. GCP reliability, 3 layers of knowledge, data streaming and more.
AWS vs. GCP reliability is wildly different
image source: Pierce Freeman, freeman.vc
Popular wisdom tells us that the big cloud providers (Amazon, Google, Microsoft if you’re generous) are relatively interchangeable - what if that’s not actually the case? With an eye towards the ongoing global microchip shortage, the author of this post ran an experiment to test GPU availability at random times of day, and the results are shocking - indeed, Amazon and Google are very much not the same in this regard. Read on to find the results, and for extra credit, extend the open source testing framework used in the experiment!
If we treat data as a product, then what is production?
How do you know what’s a throwaway, one-off report, and what’s a maintained, supported, dashboard? Different organizations have different answers to this question, but there is by no means a standard approach for data. Certainly, Mighty Canary can help tell you if a report is live, but it’s worth examining the question of what is “production”, i.e. the line in the sand beyond which the data team is responsible for maintaining what they provide. This article does just that, also asking other parallels between product and data teams, and what that can teach us.
The modern data stack is just getting started
Do you know the four core elements of the modern data stack? If not, give this a listen - it’s a podcast interview with a data-focused venture capitalist. In the first twenty minutes, she breaks down what these elements are, tools that exist in each, and challenges data leaders face in tool selection. For an insightful, concise, and up-to-date breakdown of the state of the data tool industry, look no further.
Why modern business runs on data streaming
All data moves, and some data moves faster than others. How quick is yours? This article makes a compelling case that we should all strive for as close to real-time streamed data as possible, and that slow batch processing can no longer keep pace with the speed of today’s decision-making. It also offers a helpful reminder: in the modern data stream, don’t forget to paddle.
The three critical layers of a knowledge graph
image source: Katariina Kari, Flat Pack Tech
This article, by IKEA’s Lead Ontologist, provides a simple but useful framework for understanding how a knowledge graph could be applied to any industry. A knowledge graph “represents a network of real-world entities—i.e. objects, events, situations, or concepts—and illustrates the relationship between them” (IBM). They’re often useful in revealing otherwise hard-to-identify patterns in the relationships between data, and have been applied to great effect in industries ranging from real estate to medicine - maybe your industry is next.
Enjoy this roundup?
Subscribe to our weekly newsletter and get fresh insights for analytics leaders delivered straight to your inbox.