The model thinker : what you need to know to make data work for you / Scott E. Page.

Format:

Book

Author/Creator:

Page, Scott E., author.

Language:

English

Subjects (All):

Information visualization.

Social systems--Mathematical models.

Social systems.

Social sciences--Mathematical models.

Social sciences.

Complexity (Philosophy).

Physical Description:

xiii, 427 pages : illustrations ; 25 cm

Edition:

First edition.

Place of Publication:

New York : Basic Books, [2018]

Summary:

"We confront no end of complex problems: why is inequality on the rise? Why are more and more Americans clinically obese? Does a racially diverse team make better decisions? How can we predict the outcomes of elections? At the same time, we find ourselves awash in data, be it on the opioid crisis, college admissions, genetic correlates of disease, financial transactions, or athletic performance. To confront such complexity and put that data to use, we need models: we can use linear regression to predict sales growth, or a power-law distribution to explain city sizes and book sales. Although each model offers insight, any single model will be wrong--just ask the physicist who, trying to understand barnyard animals, imagined a spherical cow. We must be able to do better. The question is simply how. In [this book], Scott E. Page gives us the answer: many-model thinking. By applying multiple diverse frameworks, we can achieve greater insights--indeed, using many models enables us to scale a hierarchy encompassing data, information, knowledge, and ultimately wisdom. Underpinning this, Page presents twenty-five broad classes of models--including models of growth, random walks, entropy, Markov chains, and many more--in a user-friendly and highly readable format, while teaching us how and when to apply them. Whether you work in science, business, government, or even literary studies, you confront complex problems, and you have more data than ever before. The Model Thinker will show how models can make that data work for you."--Jacket.

Contents:

The many-model thinker

Why model?

The science of many models

Modeling human actors

Normal distributions : the bell curve

Power-law distributions : long tails

Linear models

Concavity and convexity

Models of value and power

Network models

Broadcast, diffusion, and contagion

Entropy : modeling uncertainty

Random walks

Path dependence

Local interaction models

Lyapunov functions and equilibria

Markov models

Systems dynamics models

Threshold models with feedbacks

Spatial and hedonic choice

Game theory models times three

Models of cooperation

Collective action problems

Mechanism design

Signaling models

Models of learning

Multi-armed bandit problems

Rugged-landscape models

Opioids, inequality, and humility.

Notes:

Includes bibliographical references (pages 357-409) and index.

ISBN:

0465094627

9780465094622

OCLC:

1028523969