Python数据科学入门：全面剖析数据处理之旅

Machine

Learning

需积分: 5 51 浏览量更新于2024-07-19 收藏 4.41MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源推荐

to this differentiation, in this book I, too, use single quotes for single characters

and double quotes for character strings.

The Book Forum

The community forum for this book can be found online at the Pragmatic

Programmers web page for this book.

There you can ask questions, post

comments, and submit errata.

Another great resource for questions and answers (not specific to this book)

is the newly created Data Science Stack Exchange forum.

Your Turn

The end of each chapter features a unit called “Your Turn.” This unit has

descriptions of several projects that you may want to accomplish on your own

(or with someone you trust) to strengthen your understanding of the material.

The projects marked with a single star

are the simplest. All you need to work

on them is solid knowledge of the functions mentioned in the preceding

chapters. Expect to complete single-star projects in no more than thirty

minutes. You’ll find solutions to them in Appendix 2, Solutions to Single-Star

Projects, on page 175.

The projects marked with two stars

are hard(er). They may take you an hour

or more, depending on your programming skills and habits. Two-star projects

involve the use of intermediate data structures and well thought-out algorithms.

Finally, the three-star

***

projects are the hardest. Some of the three-star projects

may not even have a perfect solution, so don’t get desperate if you cannot find

one! Just by working on these projects, you certainly make yourself a better

programmer and a better data scientist. And if you’re an educator, think of

the three-star projects as potential mid-semester assignments.

Now, let’s get started!

Dmitry Zinoviev

dzinoviev@gmail.com

August 2016

pragprog.com/book/dzpyds

datascience.stackexchange.com

Preface • xvi

report erratum • discuss

Unit 1

Data Analysis Sequence

The steps of a typical data analysis study are generally consistent with a

general scientific discovery sequence.

Your data science discovery starts with the question to be answered and the

type of analysis to be applied. The simplest analysis type is descriptive, where

the data set is described by reporting its aggregate measures, often in a

visual form. No matter what you do next, you have to at least describe the

data! During exploratory data analysis, you try to find new relationships

between existing variables. If you have a small data sample and would like

to describe a bigger population, statistics-based inferential analysis is right

for you. A predictive analyst learns from the past to predict the future. Causal

analysis identifies variables that affect each other. Finally, mechanistic data

analysis explores exactly how one variable affects another variable.

However, your analysis is only as good as the data you use. What is the ideal

data set? What data has the answer to your question in an ideal world? By

the way, the ideal data set may not exist at all or be hard or infeasible to

obtain. Things happen, but perhaps a smaller or not so feature-rich data set

would still work?

Fortunately, getting the raw data from the web or from a database is not that

hard, and there are plenty of Python tools that assist with downloading and

deciphering it. You’ll take a closer look in Unit 2, Data Acquisition Pipeline,

on page 5.

In this imperfect world, there is no perfect data. “Dirty” data has missing

values, outliers, and other “non-standard” items. Some examples of “dirty”

data are birth dates in the future, negative ages and weights, and email

addresses not intended for use (

noreply@

). Once you obtain the raw data, the

next step is to use data-cleaning tools and your knowledge of statistics to

regularize the data set.

With clean data in your files, you then perform descriptive and exploratory

analysis. The output of this step often includes scatter plots (mentioned on

page 143), histograms, and statistical summaries (explained on page 150). They

give you a smell and sense of data—an intuition that is indispensable for

further research, especially if the data set has many dimensions.

report erratum • discuss

Data Analysis Sequence • 3

剩余217页未读，继续阅读

tommy_jiang2

粉丝: 0
资源: 8

Python数据科学入门：全面剖析数据处理之旅

Data Science Essentials in Python 无水印pdf

Python Data Science Essentials(2nd) azw3

edx智能方向python课程推荐

有什么学习numpy的平台吗

conda+install+r-essentials+in+mac

Essentials

python爬虫的参考文献

ros python

navicat premium essentials mac16.1.3汉化包

怎么查看Microsoft Security Essentials是否为最新版本

Windows Server 2016 Essentials和Windows Server 2016的区别

algorithm essentials cpp

添加Xamarin.Essentials库到Xamarin.Forms项目中

llvm essentials pdf

lux urp essentials

周克敏 essentials of robust control 课后解答

essentials for dummies

faster rcnn实现有丝分裂细胞检测python代码

ffmpeg-6.0-essentials_build.zip

最新资源