14 · Channels I — APIs

Thursday, Oct 8, 2026

Materials for this session are not published yet. They appear here before class.

Objectives

By the end of this session you can:

• Call an API and flatten JSON into a data frame.
• Keep credentials out of code with .Renviron.
• Handle rate caps and pagination.
• Snapshot raw responses so a pipeline re-runs deterministically.

What we cover

• What runs on API data: ACS demographics, FRED macro series, Census microdata — the pulls behind published papers.
• Endpoints, queries, JSON → data frames.
• Keys in .Renviron, never in code; caps and pagination.
• The snapshot convention: commit the raw response, so the pipeline re-runs even when the server changes — with saveRDS() and readRDS() as the cache format for anything that isn’t a flat table.

Verification habit. Check your counts against the publisher’s own documented totals.