news_decomp

Installation

# Install the published package.
pip install news-decomp

# Clone the repository for development.
git clone https://github.com/bank-of-england/news-decomp
cd news-decomp

# Install the package with development dependencies.
pip install -e ".[dev,docs,notebooks]"

Quick start

Run the sample-data example:

python -m examples.example_data

Calculate forecast metrics and draw charts:

python -m examples.example_analysis

Build the nowcast report:

python -m examples.example_report

Or open the interactive Marimo example:

marimo edit examples/example_data_marimo.py

The Marimo example is also published in the documentation. Regenerate its Markdown page after changing the app:

python docs/convert_notebooks.py

The decompositions table

The package consumes one long-format table. Each row records one additive component of either a forecast level or a forecast revision. Every row must satisfy the schema in src/news_decomp/schema.py; see news_decomp.md for the complete data contract.

Columns

Documentation

• News Decomposition data contract: schema, identities, and worked examples.
• API reference: public classes and methods.
• The examples/ directory contains runnable scripts for simulation, metrics, reports, and Marimo.

Data Classification

Bank of England Data Classification: OFFICIAL BLUE