No. 114 of 500 No. 11 in Machine Learning

Data Science for Business

What You Need to Know about Data Mining and Data-Analytic Thinking

By Foster Provost and Tom Fawcett

Print length
413 pages
Publisher
O'Reilly Media
Reading level
Intermediate
ISBN-13
978-1449361327
Published
September 17, 2013
Written for
Managers, consultants and analysts who work alongside data scientists and must evaluate predictive models.

Synopsis

Foster Provost and Tom Fawcett built this book from an MBA course Provost taught at NYU Stern, and its premise is that managers need data-analytic thinking more than they need algorithms. Rather than teaching code, the authors organize the field around a small set of fundamental concepts: framing a business problem as a classification, regression or clustering task, measuring how informative a variable is, recognizing overfitting, evaluating models against costs and benefits, and reasoning about similarity, probability and text. Each idea is anchored in a business case such as predicting customer churn at a wireless carrier or targeting a marketing offer, and the standard data mining process ties the chapters together.

The payoff is fluency. Readers learn to ask whether a proposed model answers the right question, whether its accuracy figure means anything for the business, and how to read lift, ROC and profit curves when comparing alternatives. The expected-value framework, which the book uses to connect model output to decisions, remains one of its most practical tools. The material predates the deep learning boom and generative AI, so it says little about neural networks or large language models, yet its principles on evaluation, generalization and overfitting apply directly to those newer systems. Some passages include equations and plots that nontechnical readers may find demanding, though no programming is required.

Scorecard

Editorial score

How it scored

Editorial score: 77.3 / 1005 weighted criteria

  • Influence30%

    24.0/30
  • Authority25%

    21.3/25
  • Practical value20%

    16.0/20
  • Currency15%

    9.0/15
  • Readability10%

    7.0/10
  • Editorial score

    77.3/100

How the rank was computed

Rank No. 114 of 500 comes from a composite of 62.1: observable signals 27.2 of 55 points plus editorial review 34.9 of 45 points.

All 11 inputs
InputPoints
Observable signals27.2/55
Reach (Amazon ratings volume)11.5/16.5
Consensus across best-of lists8.3/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted6.1/8.3
Amazon sales rank1.3/5.5
Recency0.0/2.8
Editorial review34.9/45
Author authority9.6/11.3
Influence10.8/13.5
Practical value7.2/9.0
Readability3.2/4.5
Currency4.1/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A durable bridge between machine learning concepts and business decisions, still widely taught to managers who commission and evaluate predictive models.

Strengths

  • Concept-first approach requiring no programming
  • Grounded in years of MBA classroom use
  • Evaluation chapters remain directly applicable to AI

Keep in mind

Published in 2013, it predates the deep learning boom and generative AI, and some mathematical passages may challenge readers without quantitative training.

Key takeaways

  1. Most business data problems reduce to a few task types, such as classification, regression, clustering and similarity matching, and naming the task clarifies the solution.
  2. Overfitting is the central risk in data mining, so models must be judged on holdout data rather than on the data used to build them.
  3. Accuracy alone misleads; the expected-value framework ties model predictions to the costs and benefits of each decision a business will make.
  4. Data and the capability to extract value from it should be treated as strategic assets that require deliberate, sustained investment.

Notable facts

  • Used in more than 150 colleges and universities in 26 countries and translated into 11 languages, according to the authors' book site. Source (opens in a new tab)
  • Fortune included it in its "Required Reading: The MBA List," according to NYU Stern. Source (opens in a new tab)

About the author

Foster Provost and Tom Fawcett

Foster Provost is a professor of data science and information systems at NYU Stern, served as editor-in-chief of the journal Machine Learning, received the 2020 ACM SIGKDD Test of Time Award and was on the founding teams of Dstillery and Integral Ad Science. Co-author Tom Fawcett, a machine learning researcher at GTE Laboratories and HP Labs who served as an action editor of Machine Learning, died in 2020.

Questions

Is Data Science for Business still worth reading in 2026?
Yes, for readers who need to evaluate rather than build models. Its examples predate deep learning and large language models, but its core lessons on framing problems, avoiding overfitting, testing on holdout data and tying predictions to business costs and benefits apply to any predictive system, including today's AI tools. Readers seeking coverage of neural networks or generative AI will need a newer companion text.
Do I need to know programming or math to read Data Science for Business?
No programming is required, because the book teaches concepts rather than code. It does include some equations, charts and technical terminology, especially in the chapters on model evaluation and probabilistic reasoning, so comfort with basic algebra and statistics helps. Managers without quantitative training can still follow the main arguments by focusing on the worked business examples that accompany each concept.
Who should read Data Science for Business?
The book suits MBA students, product managers, consultants and executives who commission analytics projects or work alongside data scientists, and it is widely used in university courses as an introduction to data-analytic thinking. Practicing data scientists also use it to explain their work to business colleagues. Readers wanting hands-on Python or machine learning engineering should choose a practical coding book instead.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.