FRED_Loader

A Python wrapper around the Federal Reserve FRED API that abstracts away series IDs, resampling rules, and macro scoring behind a clean, human-readable interface. Installs as fred-loader, for anyone who pulls macroeconomic data into pandas.

Install

pip install fred-loader

Or from a checkout of this repository: pip install -e ..

Quickstart

from fred_loader import Config, pull_fred

cfg = Config(filename="fred_master.csv", output_path="./data")
df = pull_fred(cfg)

Cherry-pick series categories:

from fred_loader import INFLATION, LABOR, RATES, Config, pull_fred

cfg = Config(
    filename="policy_inputs.csv",
    output_path="./data",
    series={**INFLATION, **LABOR, **RATES},
)
df = pull_fred(cfg)

Learning it

Worked notebooks live in research-kit/notebooks, alongside census-loader and otter, because the examples worth reading use this library with the others rather than on its own.

01-get-set-up assumes you have seen pandas once and gets you from nothing to a chart of something true. 03-two-sources-one-question puts FRED and Census data in one model and deals with the fact that they disagree about time.

They replaced a demo notebook that showed the library working. These show a piece of work, and the library happens to be how it gets done.

Project structure

src/fred_loader/
  __init__.py       # public API re-exports
  utils.py          # Config
  load.py           # pull_fred entry point
  series.py         # series catalog: categories and subcategories
  macro_scores.py   # scoring layer: score, available_scores
  py.typed          # PEP 561: ships the inline annotations to type checkers