#PowerBI can also make pretty dashboards like #Tableau can.
Some time ago last year, I shared my Tableau dashboard replication in Power BI.
Some people asked if the Power BI file was available somewhere.
It was not, until now!
Link to the files: https://t.co/OrlikFMSXH
Wait til your AI model starts posting memes on socials like “You’re not AI for data engineering until you dropped a prod database“
That’s when you truly know your AI knows what it’s doing
#data#ai
In college, one of my teachers taught me, "If you think you're going to use what you learn in class, you'll recognize it when you actually see it".
Yup he was right on that
#learning
Any credible research on the affect the use of AI has on human brains?
I've just been curious, because getting more things done and being “efficient” is great, but what could we be missing or losing when we rely so heavily on AI?
#ai#tech#careers
In this article, I break it down simply:
- what dimensional modeling actually is
- how star schemas work
- how it compares to other approaches
If you work with analytics data, this will make things a lot clearer: https://t.co/d2CmZAwxSX
#data#analytics hashtag#datamodeling
Good analytics starts with good data modeling.
That’s where dimensional modeling comes in:
- facts capture what happened
- dimensions provide the context
- the structure makes data easy to query and understand
🧵
It's not about the tools but it's about the tools.
If good, old spreadsheets solve your exact problem, that's what you need.
No need to make it fancy and expensive.
Call it a day, move on, and build another spreadsheet!
#data#spreadsheet
I’ve always believed in keeping data teams lean.
With AI becoming increasingly capable, I believe in that more than ever.
If you're a data person and not using AI yet, you're missing out!
#data#ai
Who hasn't heard of or used Polars by now?
If you're only using pandas, you're missing out.
Polars delivers faster performance and provides better ergonomics and consistency in its syntax.
Give my article a read that may convince you try Polars: https://t.co/fdj6xoBmRU
We're not there yet. We can't let AI just do whatever it does without any human interventions.
I'd say if your AI thing only works within the guardrails you set, then it'll be a great thing to have.
When AI did things you didn't expect? It's all on you...
#data#ai
If you're running a startup and wanting to do data science, you may want to get your data in the right structure first.
Going right into data science without a proper data set up can make your head hurt later.
And after all, maybe what you need is an analysis than data science.
It's been a little bit since I did a serious development in dbt.
I've been using the OSS SQLMesh and I'll give dbt a try again on a toy project to see how I like it and how it's different compared to SQLMesh!
If you have tried both, how do you like each tool?
#data
Missed yesterday’s issue?
I shared how to build a portable analytics stack, something that works today without locking you into heavy infra or long-term commitments.
If you didn’t catch it, you can read it here ↓
https://t.co/TlKvRdBzfA
#dataengineering
It’s live!
I just published an article on a portable analytics stack—local compute, shared state, object storage, no warehouse required.
It also shows you how dlt, ducklake, and sqlmesh work together.
👉 https://t.co/XXkNHiGcZU
#dataengineering
You don’t need a warehouse to build an analytics stack.
Tomorrow, I’ll show 20,000+ people how to build a portable analytics stack with ephemeral compute, shared state, and object storage (Thanks Daniel for the collab!)
If you want to join us, subscribe: https://t.co/nqzXv3EIJc
It always surprises me every time I work on some logic that's seemingly so simple, but it gets very complicated when you actually implement it in a model/table.
And does every business ever understand that aspect of a data job? Probably not.
We all knew this would come after us sooner or later, didn’t we?
There are plenty of other options like Airbyte, Estuary, dlt.
There is some cost in moving away from Fivetran, but it will probably be worth it in the long run.
#data#dataengineering
Managing Python projects doesn’t have to be complicated.
I shared why I switched to uv to manage Python dependencies and environments.
Yesterday, 1k people learned how to simplify their Python workflow with uv.
Did you miss the issue?
Grab it in the comment ↓