Okay, so today I’m gonna walk you through my little adventure comparing Atlas and Santos. I’m no expert, just a guy tinkering around, so bear with me.
First off, I needed a place to start. I had this project kicking around, nothing fancy, just a basic data pipeline thingy. It was sitting there, begging for some structure. That’s when I thought, “Why not give Atlas and Santos a whirl?” I’d heard whispers about them, figured it was time to get my hands dirty.
Getting Started with Atlas: I went the Atlas route first. Downloaded the whole shebang from their website. It’s a bit of a hefty download, but nothing my internet couldn’t handle. I then installed the darn thing. Followed the instructions, step-by-step, like a good little coder. It wasn’t too bad, just a bunch of next, next, next clicks. Once installed, I fired it up. The UI looked clean, not gonna lie. Started poking around, trying to figure out where everything was. Found the documentation, which was… okay. Not the best, but enough to get me started.
Setting up Santos: Next up, Santos. Now, Santos was a bit different. It felt lighter, quicker. The install was pretty straight forward. I think I used pip to install it. Boom, done. No messy GUI to deal with at first. Everything seemed code-centric. Fired up my IDE, got a new Python project going, and started writing some code to interact with Santos.
The Hands-On: Here’s where things got interesting. I decided to try and use both to describe my data pipeline. With Atlas, I was clicking around, creating entities, defining relationships, all through the GUI. It felt a bit… clunky? Maybe it’s just me, but I prefer to type stuff out. Santos, on the other hand, was all code. I defined my pipeline components as Python classes, and then used Santos to register them. I actually liked this a lot more. Felt more natural, more “coder-y”.
Data Lineage Visualization: Both tools promised data lineage visualization. With Atlas, it was built right into the GUI. Click a button, and boom, there’s your lineage. Pretty slick. Santos required a bit more work. I had to generate a graph file, then use some other tool to visualize it. A little more involved, but I had more control over the output.
The Verdict (for me): Okay, so, after playing around with both, I gotta say, I’m leaning towards Santos. It just fits my workflow better. I like the code-first approach, the flexibility, and the fact that it’s not tied to a specific GUI. Atlas is cool, but it felt a bit too enterprise-y for my little project. Your mileage may vary, of course. If you’re into GUIs and have a big, complex environment, Atlas might be a better fit. But for me, Santos is where it’s at.
That’s my two cents! Hope this helps someone out there who’s also trying to figure out which one to use. Now, back to coding!
