Day 1 : HALP
I started the technical buildout, and the reality immediately set in: before you can be a big shot dev and interpret labor-market signals, you have to earn the right to see the data.
That meant environment setup, API authentication, file structure decisions, and getting comfortable with the friction that comes with doing this as a newbie with limited coding experience.
I’m literally being fueled by my own curiosity and hubris.
I chose to work in GitHub Codespaces rather than locally. Partly for reproducibility, partly because I want this project to be inspectable end-to-end.
If I say something came from an API pull, I want that pull to exist, logged, and a proud moment in my development history, as small as it may seem.
The day was spent wiring up:
- •a Python environment,
- •secure API key handling,
- •initial calls to the Census ACS and BLS APIs,
- •a local analytics layer using DuckDB.
I hit authentication troubles, silent API errors, and confusing JSON responses that turned out to be HTML error pages (Why? … ) Anyway.
At one point, the Census API was returning “Invalid Key” while still responding with a 200 status code— didn’t really know what that meant but I knew it meant debugging so we did that and BEHOLD; there was a space where it demanded no spaces.
What is debugging in this case ? I know just from previous fiddling that I needed to see the raw responses and to double check my environment variables.
I also knew my criteria needed to be that;
- •the APIs were returning valid JSON,
- •raw responses were being saved as-is,
- •those responses could be loaded into DuckDB and queried.
API → raw data files → DuckDB tables → queryable state.
For now, I’m satisfied knowing the foundation isn’t imaginary.
Tank gad. Good Night.