Think Stats: Probability and Statistics for Programmers

Livro digital

Título:
Think Stats: Probability and Statistics for Programmers

Autor:
Allen B. Downey

Categoria:
Tecnologia > Dados

Doador:
Raffaello D. N.

Sinopse:
Statistics becomes far easier to grasp when you can make its ideas run, fail, and reveal themselves in code. Allen B. Downey begins with real survey data and statistical questions, then builds from histograms, probability mass functions, cumulative distributions, and outliers toward models that programmers can test directly. Rather than treating computation as an afterthought, the book uses Python as a way to develop intuition and expose mistaken assumptions. The chapters move through continuous distributions, probability, Bayes's theorem, operations on distributions, hypothesis testing, estimation, correlation, and linear regression. Case studies draw on health surveys, census data, marathon results, and other real sources, while exercises ask readers to simulate behavior, analyze uncertainty, and implement the methods themselves. Topics such as the central limit theorem, Monte Carlo reasoning, p-values, sampling bias, and Bayesian estimation emerge from experiments instead of arriving as disconnected formulas. This computational approach makes the material especially useful for developers, data practitioners, and students who know some programming but want a stronger statistical foundation. Readers finish with more than a catalog of techniques: they gain a repeatable way to investigate data, question apparent patterns, quantify uncertainty, and communicate conclusions with greater care. It is a practical bridge from writing code that processes numbers to reasoning responsibly about what those numbers mean.