数据科学手册：从入门到精通的完整指南

3星 · 超过75%的资源需积分: 44 25 浏览量更新于2024-07-20 6 收藏 5.86MB PDF 举报

《数据科学手册》是一本全面概述数据科学领域的综合性指南，它涵盖了数据分析、编程以及商业技能，旨在帮助读者掌握这一领域。作者以寻找独角兽般的数据科学家的难度为引子，强调了数据科学不仅仅是应用可训练技能，更需要灵活思考和理解各个领域之间的联系。本书特别注重计算机科学和软件工程的深入讲解，因为它们在数据科学家日常工作中起着核心作用。书中涵盖了经典的机器学习算法，从其数学原理到实际应用场景，突出了可视化工具在数据科学中的中心地位。同时，经典统计学也被提及，以帮助读者批判性地思考数据解读及其常见误区。此外，技术结果的清晰沟通——数据科学中常常被忽视但至关重要的技能——也得到了专门章节的阐述，所有主题都是围绕解决现实世界的数据问题展开。书中提供了大量的Python示例代码和教程，以及对大数据核心技术的介绍，包括它们的优点、局限性和如何应用于解决实际问题。作者注重实践，理论部分尽量精炼，但用直观的方式呈现，以激发读者的批判性思维和创新精神。书中的案例研究来自各行各业，为读者提供了行业实践的深入理解。此外，作者还提供了关于现代数据科学家工作现实的实用建议，包括工作流程、时间分配、处理的数据类型以及所需技能等。本书不仅适合想要从事数据分析但缺乏相应技能的人，如软件专业人员和需要了解分析的统计学家，也适合研究者和入门级研究生，他们需要学习现实世界的分析并扩展技能。 Field Cady是Maana, Inc.的首席数据科学家，他利用大数据工具解决工业问题。作者拥有斯坦福大学物理学学士学位、华盛顿大学应用数学硕士学位和卡内基梅隆大学计算机科学硕士学位。本书共分为三个部分：基础工具、还需掌握的知识和专业或高级主题，涉及广泛的议题，如数据清理、机器学习分类、自然语言处理等。《数据科学手册》于2017年首次出版，版权属于John Wiley & Sons, Inc.。本书符合版权法规定，关于如何获取使用授权的信息可访问 Wiley 的网站。Field Cady 在法律允许的范围内被确认为本书作者。本书电子版和按需印刷可能存在内容差异。本书还包括了MATLAB®商标的使用说明，这是The Math Works, Inc.的注册商标。

Preface

xviii

and to the point: ideal for somebody who needs to come up to speed quickly or

solve a problem on a tight deadline. The educational system has not yet caught

up to the demands of this new and exciting field, and my hope is that this book

will help you bridge the gap.

Field Cady

September 2016

Redmond, Washington

The Data Science Handbook, First Edition. Field Cady.

“Data science” is a very popular term these days, and it gets applied to so many

things that its meaning has become very vague. So I’d like to start this book by

giving you the definition that I use. I’ve found that this one gets right to the

heart of what sets it apart from other disciplines. Here goes:

Data science means doing analytics work that, for one reason or another,

requires a substantial amount of software engineering skills.

Sometimes, the final deliverable is the kind of thing a statistician or business

analyst might provide, but achieving that goal demands software skills that

your typical analyst simply doesn’t have. For example, a dataset might be so

large that you need to use distributed computing to analyze it or so convoluted

in its format that many lines of code are required to parse it. In many cases,

data scientists also have to write big chunks of production software that imple-

ment their analytics ideas in real time. In practice, there are usually other dif-

ferences as well. For example, data scientists usually have to extract features

from raw data, which means that they tackle very open‐ended problems such

as how to quantify the “spamminess” of an e‐mail.

It’s very hard to find people who can construct good statistical models, hack

quality software, and relate this all in a meaningful way to business problems.

It’s a lot of hats to wear! These individuals are so rare that recruiters often call

them “unicorns.”

The message of this book is that it is not only possible but also relatively

straightforward to become a “unicorn.” It’s just a question of acquiring the par-

ticular balance of skills required. Very few educational programs teach all of

those skills, which is why unicorns are rare, but that’s mostly a historical acci-

dent. It is perfectly reasonable for a single person to have the whole palette of

abilities, provided they’re willing to ignore the traditional boundaries between

different disciplines.

This book aims to teach you everything you’ll need to know to be a compe-

tent data scientist. My guess is that you’re either a computer programmer

Introduction: Becoming a Unicorn

1 Introduction: Becoming aUnicorn

looking to learn about analytics or more of a mathematician trying to bone up

on their coding. You might also be a businessperson who needs the technical

skills to answer your business questions or simply an interested layman.

Whoever you are though, this book will teach you the concepts you need.

This book is not comprehensive. Data science is too big an area for any per-

son or book to cover all of it. Besides, the field is changing so fast that any

“comprehensive” book would be out‐of‐date before it came off the presses.

Instead, I have aimed for two goals. First, I want to give a solid grounding in the

big picture of what data science is, how to go about doing it, and the founda-

tional concepts that will stand the test of time. Second, I want to give a “com-

plete” skill set, in the sense that you have the nuts‐and‐bolts knowledge to go

out and do data science work (you can code in Python, you know the libraries

to use, most of the big machine learning models, etc.), even if particular pro-

jects or companies might require that you pick up a new skill set from some-

where else.

1.1 Aren’t Data Scientists Just Overpaid

Statisticians?

Nate Silver, a statistician famous for accurate forecasting of US elections, once

famously said: “I think data scientist is a sexed‐up term for statistician.” He has

a point, but what he said is only partly true. The discipline of statistics deals

mostly with rigorous mathematical methods for solving well‐defined prob-

lems. Data scientists spend most of their time getting data into a form where

statistical methods could even be applied. This involves making sure that the

analytics problem is a good match to business objectives, extracting meaning-

ful features from the raw data and coping with any pathologies of the data or

weird edge cases. Once that heavy lifting is done, you can apply statistical tools

to get the final results, although, in practice, you often don’t even need them.

Professional statisticians need to do a certain amount of preprocessing them-

selves, but there is a massive difference in degree.

Historically, data science emerged as a field independently from statistics.

Most of the first data scientists were computer programmers or machine learn-

ing experts who were working on Big Data problems. They were analyzing

datasets of the kind that statisticians don’t touch: HTML pages, image files,

e‐mails, raw output logs of web servers, and so on. These datasets don’t fit the

mold of relational databases or statistical tools, so for decades, they were just

piling up without being analyzed. Data science came into being as a way to

finally milk them for insights.

In 20 years, I suspect that statistics, data science, and machine learning will

blur into a single discipline. The differences between them are, after all, really

31.3 How toUse This Book?

just a matter of degree and/or historical accident. But in practical terms, for the

time being, solving data science problems requires skills that a normal statisti-

cian does not have. In fact, these skills, which include extensive software engi-

neering and domain‐specific feature extraction, constitute the overwhelming

majority of the work that needs to be done. In the daily work of a data scientist,

statistics plays second fiddle.

1.2 How Is This Book Organized?

This book is organized into three sections. The first, The Stuff You’ll Always

Use, covers topics that, in my experience, you will end up using in almost any

data science project. They are core skills, which are absolutely indispensable

for data science at any level.

The first section was also written with an eye toward people who need data

science to answer a specific question but do not aspire to become full‐fledged

data scientists. If you are in this camp, then there is a good chance that Part I

of the book will give you everything you need.

The second section, Stuff You Still Need to Know, covers additional core

skills for a data scientist. Some of these, such as clustering, are so common that

they almost made it into the first section, and they could easily play a role in

any project. Others, such as natural language processing, are somewhat spe-

cialized subjects that are critical in certain domains but superfluous in others.

In my judgment, a data scientist should be conversant in all of these subjects,

even if they don’t always use them all.

The final section, Stuff That’s Good to Know, covers a variety of topics that

are optional. Some of these chapters are just expansions on topics from the first

two sections, but they give more theoretical background and discuss some

additional topics. Others are entirely new material, which does come up in data

science, but which you could go through a career without ever running into.

1.3 How to Use This Book?

This book was written with three use cases in mind:

1) You can read it cover‐to‐cover. If you do that, it should give you a self‐con-

tained course in data science that will leave you ready to tackle real prob-

lems. If you have a strong background in computer programming, or in

mathematics, then some of it will be review.

2) You can use it to come quickly up to speed on a specific subject. I have tried

to make the different chapters pretty self‐contained, especially the chapters

after the first section.

1 Introduction: Becoming aUnicorn

3) The book contains a lot of sample codes, in pieces that are large enough to

use as a starting point for your own projects.

1.4 Why Is It All in Python

™

, Anyway?

The example code in this book is all in Python, except for a few domain‐ specific

languages such as SQL. My goal isn’t to push you to use Python; there are lots

of good tools out there, and you can use whichever ones you want.

However, I wanted to use one language for all of my examples. This keeps the

book readable, and it also lets readers follow the whole book while only know-

ing one language. Of the various languages available, there are two reasons why

I chose Python:

1) Python is the most popular language for data scientists. R is its only major

competitor, at least when it comes to free tools. I have used both exten-

sively, and I think that Python is flat‐out better (except for some obscure

statistics packages that have been written in R and that are rarely needed

anyway).

2) I like to say that for any task, Python is the second‐best language. It’s a jack‐

of‐all‐trades. If you only need to worry about statistics, or numerical com-

putation, or web parsing, then there are better options out there. But if you

need to do all of these things within a single project, then Python is your

best option. Since data science is so inherently multidisciplinary, this makes

it a perfect fit.

As a note of advice, it is much better to be proficient in one language, to the

point where you can reliably churn out code that is of high quality, than to be

mediocre at several.

1.5 Example Code and Datasets

This book is rich in example code, in fairly long chunks. This was done for two

reasons:

1) As a data scientist, you need to be able to read longish pieces of code. This

is a nonoptional skill, and if you aren’t used to it, then this will give you a

chance to practice.

2) I wanted to make it easier for you to poach the code from this book, if you

feel so inclined.

You can do whatever you want with the code, with or without attribution. I

release it into the public domain in the hope that it can give some people a small

leg up. You can find it on my GitHub page at www.github.com/field-cady.

剩余394页未读，继续阅读

ramissue

粉丝: 354
资源: 1487

数据科学手册：从入门到精通的完整指南

大数据科学手册

Data Science Programming in Python 无水印pdf

Data Science Fundamentals for Python and MongoDB 无水印原版pdf

reliability data handbook *.pdf

有关python大数据分析技术的文献及其作者和出处

python data science handbook csdn

能推荐一些学习python数据分析的书籍、网站、论坛吗

handbook of antenna technologies 2016.pdf

数据爬取与可视化分析参考文献

国外文本挖掘研究现状和参考文献

最新资源