Modelling Paradigms

Modelling paradigms for data-driven science: learning from example

Model-building praxis in data-driven science is quite distinctive. As well as predicting and explaining outcomes, models are often used as containers for structuring information and learning within the data space. We will consider the ways in which models are incorporated into interactive data analysis by asking:

• How relevant models are identified and selected;
• How they are developed and refined; and
• The processes through which they are analysed and evaluated.

From here, we will document what is distinctive about model building in data-driven science.

Related work (challenge team)

• Beecham, R. and Lovelace, R. (2023) A Framework for Inserting Visually Supported Inferences into Geographical Analysis Workflow: Application to Road Safety Research, Geographical Analysis, 55(3): 345–366.

• Cook, D., Reid, N., and Tanaka, E. (2021) The Foundation Is Available for Thinking About Data Visualization Inferentially. Harvard Data Science Review, 3(3).

• Franklin, R. (2023) Quantitative methods II: Big theory. Progress in Human Geography, 47(1): 178-186.

• Gelman, A. (2004) “Exploratory Data Analysis for Complex Models.” Journal of Computational and Graphical Statistics, 13: 755–79.

• Hullman, J., and Gelman, A. (2021) Designing for Interactive Exploratory Data Analysis Requires Theories of Graphical Inference. Harvard Data Science Review, 3(3).

• Wolf, L.J. (2022) “The Right to Rule by Thumb: A Comment on Epistemology in A Route Map for Successful Applications of Geographically-Weighted Regression, Geographical Analysis, 55(1): 184–190.

• Wood, J., D. Badawood, J. Dykes, and A. Slingsby. (2011). “BallotMaps: Detecting Name Bias in Alphabetically Ordered Ballot Papers.” IEEE Transactions on Visualization and Computer Graphics, 17: 2384–91.