Data News — Week 13
Data News #13 — I was sick so I wrote stuff with no point this week.
I've been sick the whole week and writing this week edition feels a bit weird from me. What I like about the weekly newsletter routine is to discover people, products, stuff, crazy ideas, etc. But this week as my brain wasn't able to function at full potential navigating through all articles felt different. So it's gonna be different.
Benn, as usual, achieved my mental health. First because each time I read him I ask myself why people read my primary school English while there is him, Tristan, Sarah and the whole Substack data clique weekly distilling words about data. Secondly because he wrote about Data Council Austin and how it was a data professionals only party auto-satisfying themselves. Other people felt the same way and the critics were saying, for the first time, business users were forgotten. But we do data for business — do we? — not only for the sake of building tools and platform.
I'm also concerned regarding the future of data conferences, this is a side effect of the unbundling, as we have more and more companies we'll see a fight between conferences, often marketing conferences. We will be left with our Panini album trying to collect ideas on everything and everywhere on the globe. As an European I've always felt jet-lagged by big conferences, American people flying over to show us how they do stuff in order for us to copy paste because it's the way. It is deeply integrated in me. This is something I'd love to help change in the coming times.
Where am I going?
The whole point of this newsletter from the beginning has been to save ideas and links I like to nourish my mind and to shine when I speak with data people. So the moto had always been: write something you want to read. As an independent I also want to keep a neutrality, I mean I don't want to favour a product over another one or have an unbalanced edito. Because if I want you to trust and support me I have to show you that I'm not "sold". To be honest when we look at the numbers it's unbalanced, I've shared more stuff about dbt and Airflow than anything. But there are reasons, I really like these two products and I sell my expertise on those on the other side.
My expertise, something I've kept apart from the data news. For the last 2 years I've done consulting for French organisations — public, private, startups. I really like this part of my work but it's hard to conciliate content creation and client facing activities, especially when you have 5 clients at the same time and that you're alone. Anyway I'll bring new stuff soon.
Recently, for instance, I helped a startup — a bank for kids — to recruit their 2 first data profiles, DE and DA, and it was an awesome experience. I was running technical interviews. I enjoy myself the most when I see the value I can bring to a company and here being like a dating agency trying to match candidates and company was perfect.
And the Fast News ⚡️
And because you subscribed for data news here some articles I found interesting this week.
- The Soldiers, Rogues, and Mages of Data Teams — Jesse is saying that Data Teams are like RPGs. I can't wait to see this on companies career paths. I want to be a dwarf.
- Comparing Flyte and Airflow — explanations from Lyft team after Spotify announcement last week.
- S3 is not a Backup — Don't forget to build a complete strategy, S3 should not be your only answer
- Computer Vision pipeline with Kubernetes — namR, a Paris based company, tries to infer parcels with aerial imagery with Kubernetes as orchestrator.
- Features & Labels team developed a way to run Python scripts after a dbt SQL model execution by specifying it in the YAML tags.
- Beat shared their dbt setup to provide company data marts on top of Trino by using Kubernetes
A pastel note 🎨
If you are still struggling to understand how Random Forest algorithm works this scroll to animate visualisation will help you a lot.
To finish this late edition of the data news different than what you are used to — next week gonna be as usual don't worry — Yesterday on April 1st people started the #30DayChartChallenge on Twitter. Justin just did a heatmap representing Ballon d'Or winners by nationality and club, I really like the color palette and the way he annotated the visualisation.
Join the newsletter to receive the latest updates in your inbox.