arXiv subjects

Thomas A. Grossman

Publications and source records attributed to Thomas A. Grossman.

12 recordsLinked to original sources

arXiv2026

Spreadsheet Modeling Experiments Using GPTs on Small Problem Statements and the Wall Task

This paper investigates how GPT-based tools can assist in building reusable analytical spreadsheet models. After a screening, we evaluate five GPT extensions and select Excel AI by pulsrai.com for detailed testing. Through structured experiments on simple problem statements, we assess Excel AI's performance against the ERFR criteria (each input in a cell; cell formulas; no hardwired numbers; labels; accurate). Results show that while Excel AI can produce well-structured models, it is inconsistent and often non-reproducible. We identify two central challenges - "the problem of confidence" and "the problem of workflow" - which highlight the need for skilled users to verify and adapt GPT-generated spreadsheets. Though GPTs show promise for generating draft models that may reduce development time or lower skill requirements, current tools remain unreliable for professional use. We conclude with recommendations for future research into prompt engineering, reproducibility, and larger-scale modeling tasks.

cs.SE↗

arXiv2023

A Use Case-Engineering Resources Taxonomy for Analytical Spreadsheet Models

This paper presents a taxonomy for analytical spreadsheet models. It considers both the use case that a spreadsheet is meant to serve, and the engineering resources devoted to its development. We extend a previous three-type taxonomy, to identify nine types of spreadsheet models, that encompass the many analytical spreadsheet models seen in the literature. We connect disparate research literature to distinguish between an "analytical solution" and an "industrial-quality analytical spreadsheet model". We explore the nature of each of the nine types, propose definitions for some, relate them to the literature, and hypothesize on how they might arise. The taxonomy aids in identifying where various spreadsheet development guidelines are most useful, provides a lens for viewing spreadsheet errors and risk, and offers a structure for understanding how spreadsheets change over time. This taxonomy opens the door to many interesting research questions, including refinements to itself.

cs.SE↗

arXiv2018

Alternative Spreadsheet Model Designs for an Operations Management Model Embedded in a Periodic Business Process

We present a widely-used operations management model used in supply and distribution planning, that is typically embedded in a periodic business process that necessitates model modification and reuse. We consider three alternative spreadsheet implementations, a data-driven design, a canonical (textbook) design, and a novel (table-driven) technical design. We evaluate each regarding suitability for accuracy, modification, analysis, and transfer. We consider the degree of training and technical sophistication required to utilize each design. The data-driven design provides insight into poor spreadsheet practices by naïve modelers. The technical design can be modified for new data and new structural elements without manual writing or editing of cell formulas, thus speeding modification and reducing risk of error. The technical design has potential for use with other classes of models. We identify opportunities for future research.

cs.SE↗

arXiv2011

Towards Evaluating the Quality of a Spreadsheet: The Case of the Analytical Spreadsheet Model

We consider the challenge of creating guidelines to evaluate the quality of a spreadsheet model. We suggest four principles. First, state the domain-the spreadsheets to which the guidelines apply. Second, distinguish between the process by which a spreadsheet is constructed from the resulting spreadsheet artifact. Third, guidelines should be written in terms of the artifact, independent of the process. Fourth, the meaning of "quality" must be defined. We illustrate these principles with an example. We define the domain of "analytical spreadsheet models", which are used in business, finance, engineering, and science. We propose for discussion a framework and terminology for evaluating the quality of analytical spreadsheet models. This framework categorizes and generalizes the findings of previous work on the more narrow domain of financial spreadsheet models. We suggest that the ultimate goal is a set of guidelines for an evaluator, and a checklist for a developer.

cs.SE↗

arXiv2010

Spreadsheets Grow Up: Three Spreadsheet Engineering Methodologies for Large Financial Planning Models

Many large financial planning models are written in a spreadsheet programming language (usually Microsoft Excel) and deployed as a spreadsheet application. Three groups, FAST Alliance, Operis Group, and BPM Analytics (under the name "Spreadsheet Standards Review Board") have independently promulgated standardized processes for efficiently building such models. These spreadsheet engineering methodologies provide detailed guidance on design, construction process, and quality control. We summarize and compare these methodologies. They share many design practices, and standardized, mechanistic procedures to construct spreadsheets. We learned that a written book or standards document is by itself insufficient to understand a methodology. These methodologies represent a professionalization of spreadsheet programming, and can provide a means to debug a spreadsheet that contains errors. We find credible the assertion that these spreadsheet engineering methodologies provide enhanced productivity, accuracy and maintainability for large financial planning models

cs.SE↗

arXiv2009

The Lookup Technique to Replace Nested-IF Formulas in Spreadsheet Programming

Spreadsheet programmers often implement contingent logic using a nested-IF formula even though this technique is difficult to test and audit and is believed to be risky. We interpret the programming of contingent logic in spreadsheets in the context of traditional computer programming. We investigate the "lookup technique" as an alternative to nested-IF formulas, describe its benefits for testing and auditing, and define its limitations. The lookup technique employs four distinct principles: 1) make logical tests visible; 2) make outcomes visible; 3) make logical structure visible; and 4) replace a multi-function nested-IF formula with a single-function lookup formula. It can be used only for certain simple contingent logic. We describe how the principles can be applied in more complex situations, and suggest avenues for further research.

cs.SE↗

arXiv2008

A Primer on Spreadsheet Analytics

This paper provides guidance to an analyst who wants to extract insight from a spreadsheet model. It discusses the terminology of spreadsheet analytics, how to prepare a spreadsheet model for analysis, and a hierarchy of analytical techniques. These techniques include sensitivity analysis, tornado charts,and backsolving (or goal-seeking). This paper presents native-Excel approaches for automating these techniques, and discusses add-ins that are even more efficient. Spreadsheet optimization and spreadsheet Monte Carlo simulation are briefly discussed. The paper concludes by calling for empirical research, and describing desired features spreadsheet sensitivity analysis and spreadsheet optimization add-ins.

cs.SE↗

arXiv2008

Accuracy in Spreadsheet Modelling Systems

Accuracy in spreadsheet modelling systems can be reduced due to difficulties with the inputs, the model itself, or the spreadsheet implementation of the model. When the "true" outputs from the system are unknowable, accuracy is evaluated subjectively. Less than perfect accuracy can be acceptable depending on the purpose of the model, problems with inputs, or resource constraints. Users build modelling systems iteratively, and choose to allocate limited resources to the inputs, the model, the spreadsheet implementation, and to employing the system for business analysis. When making these choices, users can suffer from expectation bias and diagnosis bias. Existing research results tend to focus on errors in the spreadsheet implementation. Because industry has tolerance for system inaccuracy, errors in spreadsheet implementations may not be a serious concern. Spreadsheet productivity may be of more interest.

cs.HC↗

arXiv2008

Research Strategy and Scoping Survey on Spreadsheet Practices

We propose a research strategy for creating and deploying prescriptive recommendations for spreadsheet practice. Empirical data on usage can be used to create a taxonomy of spreadsheet classes. Within each class, existing practices and ideal practices can he combined into proposed best practices for deployment. As a first step we propose a scoping survey to gather non-anecdotal data on spreadsheet usage. The scoping survey will interview people who develop spreadsheets. We will investigate the determinants of spreadsheet importance, identify current industry practices, and document existing standards for creation and use of spreadsheets. The survey will provide insight into user attributes, spreadsheet importance, and current practices. Results will be valuable in themselves, and will guide future empirical research.

cs.HC↗

arXiv2008

A Paradigm for Spreadsheet Engineering Methodologies

Spreadsheet engineering methodologies are diverse and sometimes contradictory. It is difficult for spreadsheet developers to identify a spreadsheet engineering methodology that is appropriate for their class of spreadsheet, with its unique combination of goals, type of problem, and available time and resources. There is a lack of well-organized, proven methodologies with known costs and benefits for well-defined spreadsheet classes. It is difficult to compare and critically evaluate methodologies. We present a paradigm for organizing and interpreting spreadsheet engineering recommendations. It systematically addresses the myriad choices made when developing a spreadsheet, and explicitly considers resource constraints and other development parameters. This paradigm provides a framework for evaluation, comparison, and selection of methodologies, and a list of essential elements for developers or codifiers of new methodologies. This paradigm identifies gaps in our knowledge that merit further research.

cs.HC↗

arXiv2008

Source Code Protection for Applications Written in Microsoft Excel and Google Spreadsheet

Spreadsheets are used to develop application software that is distributed to users. Unfortunately, the users often have the ability to change the programming statements ("source code") of the spreadsheet application. This causes a host of problems. By critically examining the suitability of spreadsheet computer programming languages for application development, six "application development features" are identified, with source code protection being the most important. We investigate the status of these features and discuss how they might be implemented in the dominant Microsoft Excel spreadsheet and in the new Google Spreadsheet. Although Google Spreadsheet currently provides no source code control, its web-centric delivery model offers technical advantages for future provision of a rich set of features. Excel has a number of tools that can be combined to provide "pretty good protection" of source code, but weak passwords reduce its robustness. User access to Excel source code must be considered a programmer choice rather than an attribute of the spreadsheet.

cs.SE↗

arXiv2007

Spreadsheet Engineering: A Research Framework

Spreadsheet engineering adapts the lessons of software engineering to spreadsheets, providing eight principles as a framework for organizing spreadsheet programming recommendations. Spreadsheets raise issues inadequately addressed by software engineering. Spreadsheets are a powerful modeling language, allowing strategic rapid model change, and enabling exploratory modeling. Spreadsheets users learn slowly with experience because they focus on the problem domain not programming. The heterogeneity of spreadsheet users requires a taxonomy to guide recommendations. Deployment of best practices is difficult and merits research.

cs.SE↗