All Hands on Data #24
Spooky season is finally behind us. We hope you (and your buckets of candy) enjoy this week's set of articles from our team!
The Eternal Suffering of Data Practitioners: Part 1
I started my data career after teaching as a one-man data team. I was overwhelmed and honestly discouraged as my stakeholders wanted more and more, and I could not give it to them. Even though I am a few years removed from that experience, I still get anxious thinking about it. Pedram’s article breathed a breath of fresh air into that situation. If you feel overwhelmed about your role, check this article out. - Steven Johnson
Reinventing the Wheel of Data Activation
Sarah argues that reverse ETL tools are just a new flavor of tools like Zapier and Make. If you want to act on your data, for right now, there always has to be a middleman operation. But does the future indicate that will continue to hold true? - Blake Burch
Data Swamps and The Tragedy of The Commons
Barr makes an interesting point about the faulty thinking around “data is everyone’s responsibility.” I see similar issues in the same line with “QA is everyone’s responsibility.” Bar does a nice job detailing 4 strategies to implement guardrails into a data teams process. - Jon Davidson
How to Make Python Code Run Incredibly Fast
Python is often knocked for being slow, but many times there are simple optimizations that you can make that will speed things in a major way. - Wes Poulsen
Dark Data Can Be The Next Dark Horse of Data Analytics
There’s a lot of unstructured and unindexed data out there and the overwhelming majority of it currently isn’t being utilized in any machine learning training (e.g. it’s private, behind firewalls, stored in inaccessible locations, etc). Being able to access this and feed it into models in a usable way could unlock a great deal more value from machine learning models assisting business operations. - John Forstmeier
Maturity of Machine Learning Systems
If your Machine Learning models are not being monitored, tested, and improved frequently, they can cause more damage than good. Reshytko explains the importance of measuring the maturity of your ML systems by employing a maturity framework. Incorporate Continuous Integration, Continuous Delivery, and Continuous Training into your ML systems to increase robustness and maturity. - Katt Baum
From 3D Contour Plots to AI-Generated Art
I recently posted about making music with your data. This feels like an appropriate next step in the series - making AI generated art with Python. - Joseph McDermott