Best Data Analysis Books: 35 Picks Recommended by CEOs

Hans Rosling and his co-authors test how badly people misread global data on poverty, population and schooling, and trace the errors to ten instincts. Bill Gates calls its framework a breakthrough, and Aidan Connolly recommends it for the real world of statistics.

Damian Mogavero, founder of the restaurant software company Avero, shows how leading chefs and restaurateurs use data to shape recipes, wine lists, presentation and server training. Gary Loveman says the analytics work brought great value to Caesars.

Darrell Huff, a journalist rather than a statistician, catalogs the ways figures get bent: skewed samples, loaded interview methods, truncated chart axes and pictograms that misstate scale. Bill Gates finds it more relevant than ever, and Morgan Housel says it leaves readers skeptical of nearly everything.

Ben Lindbergh and Travis Sawchik report on how baseball teams now use data, high speed cameras and new analytical tools to develop players rather than only evaluate them. Nate Silver says it shows that success owes more to a scientific approach than to natural talent.

Mark Jeffery explains 15 metrics for measuring what marketing spending returns, from brand awareness to online campaigns, with worked examples and templates for calculating return on marketing investment. Jeff Bezos has recommended it, a fit with Amazon’s rule that every claim be backed by data.

Morten Jerven examines how African statistical offices actually produce GDP figures and finds the numbers that donors rely on far less reliable than they look, arguing that no econometric technique can rescue bad data. Bill Gates finds his case against many trusted measurements strong.

Nate Silver, the forecaster behind FiveThirtyEight, examines why most predictions fail in weather, elections, poker and markets, and argues for Bayesian updating as fresh data arrives. Bill Gates notes that its subject reaches well beyond politics.

Jeffrey M. Wooldridge’s textbook shows empirical researchers how to use econometric methods on questions of business, policy evaluation and forecasting, with chapters organized by the kind of data at hand. Eric Weinstein includes it among the books he chose to keep.

Edward R. Tufte collects strong displays of complex data, from maps and timetables to statistical graphics, and draws practical lessons on small multiples, layering, color and narrative. Kevin Rose regards Tufte as the leading authority on making hard data easy to read in charts and graphs.

Steve Lohr, a New York Times technology reporter, follows the rise of big data, from young data science entrepreneurs to IBM, and asks what happens when decisions rest more on analysis than on intuition. Thomas Goetz credits Lohr with separating the real power of data from its limits.

Alistair Croll and Ben Yoskovitz show founders which numbers matter at each stage of a startup, how to read and report them, and why one metric should dominate at a time, drawing on more than 30 case studies. Ola Olusoga recommends it.

Michael Lewis follows Oakland general manager Billy Beane as he uses sabermetrics, statistical analysis that values on-base and slugging percentage over batting average, to compete against far richer teams. Scott Keyes recommends it among Lewis’s books.

Avinash Kaushik lays out a framework for turning web analytics into an actionable strategy, covering analytical technique, measurement of social and multichannel campaigns, experimentation and customer listening. Raluca Radu, who works in digital marketing, recommends it alongside his earlier An Hour A Day.

Avinash Kaushik argues that clickstream numbers alone fall short and sets out a week by week program for web analysis, from foundational metrics and internal site search to testing, surveys and usability studies. Raluca Radu has it on her list of business books.

Peter E. Kennedy explains the main ideas of econometrics, from time series and bootstrapping to instrumental variables, with intuition in place of heavy notation. Nassim Nicholas Taleb says no text builds statistical intuition better, and he has kept reading new editions over his career.

David Freedman teaches readers to follow and build statistical models through published health and social science studies, testing whether each model really connects to what it claims to describe. Nassim Nicholas Taleb praises its clarity and its use of real cases to expose the limits of statistics.

Trevor Hastie, Robert Tibshirani and Jerome Friedman place machine learning and data mining inside a single statistical framework, covering prediction, classification trees, boosting, random forests and clustering. Nassim Nicholas Taleb uses it as his reference for the inner workings of machine learning.

Mainly an economics book about incentives, Steven D. Levitt and Stephen J. Dubner’s work shows data analysis applied to odd questions, from cheating teachers and sumo wrestlers to what baby names reveal. Malcolm Gladwell credits it with inventing a whole genre.

Mainly a sales book, Mark Roberge’s account shows how an MIT-trained engineer scaled HubSpot’s sales team on metrics and process, with formulas for hiring, training, the sales process and lead flow. Anton Sepetov praises its mathematical approach to selling.

Sean Ellis and Morgan Brown write mainly about growth strategy, built on cross-functional teams running quick cycles of testing and iteration to win, keep and monetize customers. Nir Eyal credits them with getting product, data and marketing people to work as one team.

A textbook mainly about deep learning, it starts from the probability, linear algebra and numerical computation that data work rests on, then covers the networks used in industry. Elon Musk calls it the one comprehensive text on the subject, and Nassim Nicholas Taleb praises its clear math.

Jordan Ellenberg writes mainly about mathematical thinking, but his section on inference covers regression to the mean, Berkson’s paradox, p-hacking and the replication crisis, all traps in reading data. Bill Gates finds the writing funny and accessible for a math book.

Pedro Domingos explains machine learning, in which computers find knowledge in data, by sorting the field into five schools, from Bayesian probability to evolutionary computation. Mark Cuban says it persuaded him that the moment for AI had arrived.

Philip E. Tetlock’s study of political forecasting measures expert predictions against informed laypeople and simple trend extrapolations, a lesson in scoring claims against outcomes. Marc Andreessen argues that pundits should publish their track records, and Patrick Collison lists it on his bookshelf.

Jeremy N. Smith profiles Christopher Murray, the doctor and economist behind the Global Burden of Disease studies, whose premise is that health gains depend on first measuring illness precisely. Mainly a biography, it shows a vast data project ranking what makes people sick.

Data scientist Hannah Ritchie reads the evidence on soils, oceans, plastics and food from a wider angle and finds more progress than doomsday headlines suggest, a case study in handling data with care. Bill Gates says anyone discussing climate change should read it.

Mainly a strategy book, Marco Iansiti and Karim R. Lakhani’s study argues that data, analytics and AI are changing what a firm is, with Amazon, Ant Financial and Microsoft as examples of companies rebuilt around data. Mark Cuban recommends it.

Peter L. Bernstein traces how probability theory, statistics and risk management developed, arguing that tools for measuring risk freed people from relying on oracles and soothsayers. Marc Andreessen has it on his career planning reading list.

Anil Ananthaswamy explains the mathematics under machine learning, showing that it rests on old ideas such as linear algebra and calculus, groundwork for anyone curious how models learn from data. Tyler Cowen gives it a claim on his best non-fiction list, while warning that the math is demanding.

Ash Maurya writes mainly about turning a startup idea into a viable business, and his tools are metrics: traction measures, a quick viability estimate and numbers for tracking company health and reporting to investors. Eric Ries calls it the follow-up to Running Lean.

Michael D. Smith and Rahul Telang write mainly about the entertainment business, showing how Netflix used subscriber data to back House of Cards and why studios and labels now need to follow the data. Marvin Liao has it on his reading list.

Leonard Mlodinow’s book is about randomness in everyday life, and it explains why polls, wine ratings and school grades deserve less trust, along with ideas such as regression toward the mean and why large samples settle down. Gabriel Coarnă says its Monty Hall section changed how he thinks about probability.

Mildred L. Patten and Michelle Newhart introduce quantitative and qualitative research methods through short, self-contained topics with exercises, groundwork for designing studies and judging evidence. Denise Morris Kipnis recommends it to young people entering her field.

Philip E. Tetlock and Daniel Gardner study prediction rather than data analysis as such, drawing on a forecasting tournament to show that the best forecasters weigh varied evidence, think in probabilities and keep score. Daniel Kahneman calls it a rare book that leaves readers smarter.

Albert-László Barabási writes mainly about success, but his findings come from large scale data analysis by his research team on why performance often goes unrewarded and how networks shape recognition. Nassim Nicholas Taleb calls it an imperative project that applies modern science to randomness and success.