Mitigating Spreadsheet Model Risk with Python Open Source Infrastructure

TL;DR

This paper proposes an open source Python-based infrastructure to help spreadsheet professionals develop reproducible audit tools, reducing errors and improving model reliability by establishing clear testing protocols.

Contribution

It introduces a framework using Python packages for creating audit tools and model 'oracles' to enhance spreadsheet model testing and validation.

Findings

Enables development of reproducible audit tools for spreadsheets

Facilitates testing of spreadsheet calculations against defined 'oracles'

Promotes use of open source Python packages for model validation

Abstract

Across an aggregation of EuSpRIG presentation papers, two maxims hold true: spreadsheets models are akin to software, yet spreadsheet developers are not software engineers. As such, the lack of traditional software engineering tools and protocols invites a higher rate of error in the end result. This paper lays ground work for spreadsheet modelling professionals to develop reproducible audit tools using freely available, open source packages built with the Python programming language, enabling stakeholders to develop clearly defined model "oracles" with which to test and audit spreadsheet calculations against.