Tools for Value-Driven Design

Tools intended to help with modelling decisions in a value centric design process. The intent is to keep this as generic as possible, as some of this decision modelling is suited to generic decision-making, non-design activities with a little massaging.

Features

• Concept Design Analysis (CODA) method implementation
• Requirements weighting with a Binary Weighting Matrix
• Programmatic or Spreadsheet based model creation (via Excel workbooks or Google Sheets).
• Native JSON serialisation for both CODA models and Binary Weighting Matrices (human-readable, git-diff-friendly).
• Command-line interface for quick model inspection and interactive weighting

Install

pip install vdd

CLI

vdd ships with a command-line interface for the two core workflows.

Inspect a CODA model

Load a CODA model from an Excel file and print the overall design merit alongside a per-requirement satisfaction summary:

vdd coda path/to/model.xlsx

By default the compact Excel format is assumed. Pass --parser full for the standard format:

vdd coda path/to/model.xlsx --parser full

Weight requirements interactively

Step through a pairwise comparison of requirements and print the resulting normalised scores:

vdd requirements weight "Lightweight" "Stiff" "Durable"

Questions are shuffled by default. Pass --no-shuffle to work through them in a fixed order.

JSON model i/o

Both CODA models and Binary Weighting Matrices serialise to and from a flat, human-readable JSON format. This is handy for version-controlling models or moving them between the two tools.

from vdd.coda.models import CODA
from vdd.requirements.models import BinWM

# Weight requirements with a binary weighting matrix, then persist it.
bwm = BinWM('Stiffness', 'Friction', 'Weight')
bwm.prompt()
bwm.to_json('weights.json')

# Seed a CODA model directly from the weighting matrix scores.
coda = CODA()
coda.add_requirements_from(bwm)

# Round-trip a CODA model through JSON.
coda.to_json('model.json')
reloaded = CODA.read_json('model.json')

to_json() returns a string when no path is given. to_dict() / from_dict() expose the same data as plain-python dicts. The BinWM format ({"requirements": [...], "binary_matrix": [[...]]}) is the same shape used by the bundled example fixtures.

Documentation

Currently just stored in the repo.

• Using Google Sheets for requirements matrices
• tbc

Development

In the repository root:

uv sync

Releases

Managed by Release Please with auto-versioning. Changes to default branch will be accumulated in a release PR based on Conventional Commits. Merging the release PR will automatically publish the versioned package to PyPI.

Roadmap

• Model sets for comparative work (rather than a single set of characteristic parameter values)
• Improved visualisation
• Export CODA models to Excel template
• House of Quality style requirement/characteristic weighting
• Pandas everywhere (v1.x)

References

Based on my own degree notes and open access literature:

• M.H. Eres et al, 2014. Mapping Customer Needs to Engineering Characteristics: An Aerospace Perspective for Conceptual Design - Journal of Engineering Design pp. 1-24 http://eprints.soton.ac.uk/id/eprint/361442