Data Notes
About

If AI takes over the routine, we get to create.

So I started this blog. Here I run data studies with the help of AI agents: they collect sources, clean the data and do the heavy lifting, and I ask the questions, check the numbers and decide what the results mean.

Maybe the findings will be useful to you. Read along, and if you have questions or ideas for the next study, message me on social media.

— Illia Hrebenko

2

studies published

5

topics tagged

100%

of the data linked

How every study works

  1. 1
    A question. Something concrete, like which CPU brand gives more performance per dollar.
  2. 2
    Data, collected by agents. Every number keeps a link to its source, and raw pages are saved.
  3. 3
    Analysis in Python. Plain scripts, so each result can be re-run and checked.
  4. 4
    A human check. I review the method, the outliers and the claims before anything is published.
  5. 5
    A live follow-up. When a study is about listed companies, the post shows their share price on the day it was published next to the live price, so you can see how the story plays out.

Get in touch