The Art of Data Science

by Roger D. Peng, Elizabeth Matsui

This book describes, simply and in general terms, the process of analyzing data. The authors have extensive experience both managing data analysts and conducting their own data analyses, and have carefully observed what produces coherent results and what fails to produce useful insights into data. This book is a distillation of their experience in a format that is applicable to both practitioners and managers in data science.

Genres: Technology, Science, Nonfiction, Computer Science, Business, Mathematics, Reference

170 pages, Paperback

First published September 5, 2015

Reviews

[Alex · Rating 3 out of 5]

I'm somewhat ambivalent regarding this book -- I very much appreciate the pragmatic writing style, and there are some genuinely useful pieces of advice contained within. However, the target audience seems ambiguous. The best fit seems to be folks who are intending to take the JHU data science classes, and retrospectively this looks like it would be a very handy companion guide to the course. Having taken them, however, along with a number of other statistics/data analysis type classes, much of the content seems too cursory in review.

[Ahmad · Rating 5 out of 5]

The authors share their experience in data analysis and the steps they propose seems necessary for a neat data analysis. I think I should re-read this book throughout my future data analysis projects.

[Ayas · Rating 3 out of 5]

I wouldn't recommend this to someone who isn't an advanced or works as a data scientist.
Although I am pretty sure it would be too good for some one who Is/Does.

The book explains the untaught process, or lets say the unspoken process of a data scientist's job.
It explains the things you wouldn't read in a book of statistics, it explains the thought process taking place in a data scientist's mind.
It discusses the major steps taken to complete a task and how to judge every step you take.
It introduces the so called Epicycle and how it works.
But I will have to say this, If you don't work as a data scientist I think part of this book will confuse you or even scare you away. Some parts are easily understood through the first half of the book at least, but there are lots of unclear stuff and Jargons that will leave you really confused. Some examples where legit others again need way more advanced knowledge not because they are hard to understand but i felt they were a bit unclear or directed to let's say a "data scientist"
Again I am only saying that If you think this book is an introduction to data science in any kind then you might get a little disappointed.

[Alex · Rating 4 out of 5]

3.5 stars rounded up. in essence a more polished version of executive data science, and a lot my review of executive data science also applies to this book, although 'the art of data science' is more ready for public consumption. still is based on a procedural frame without much explanation for why you'd want to take the steps in the procedure largely. i think this is the kind of thing that lost undergrads will appreciate a lot, but it's pretty sloppy about estimation formalisms and this will leave those with advanced training somewhat disappointed imo. additionally the framing alternated between "the audience/context/system receiving your analytic product is constantly present and influencing your choices" and "you touch base with the audience/context/system receiving your analytic product intermittently but basically are on your own and just need to produce correct work," and i feel that there is a lot more meaningful exploration that could be done in this vein of thinking

[danielle; ▵ · Rating 3 out of 5]

3.5. A good overview for someone with some practical experience in analytics who wants to better understand the data analysis workflow.

[Aina · Rating 4 out of 5]

This excellent book takes you through each step of a typical data science project giving general advice, warning about common mistakes and giving many practical examples (including real industrial data science projects) to illustrate each of the points. It helps to build a mind map of options available at a data scientist’s disposal during each of the project stages. Definitely a book to read multiple times.

[Javad Ebadi · Rating 4 out of 5]

This book is very good to obtain a big picture about what is data analysis. The most important lesson that I learnt from this book was that a data analysis starts with a question, not with the data and at the end of the day it leads to another (better) question. The challenges in data analysis processes are described very well. Various types of question and types of data analysis are explained.

The book is very good for beginners, also it can be used by sophisticated data scientists.

[Lolo · Rating 2 out of 5]

I work with data, but I guess this book isn't for me.

I got the feeling that is was addressed to the experienced data scientist and not someone that wants to understand a little bit more about it. It seems that the author focuses more on the process and the logistics of the day-to-day tasks of a data scientist rather than the field of data science.

The book was interesting and well written but didn't really answer my questions. For sure I didn't learn the "art" of data science.

[Ms_prue · Rating 5 out of 5]

This book equipped me to answer all the questions I have in my data analyst life, specifically the "why am I here?" and "what is my purpose?" type ones. In my work, there are a lot of technical resources for data analysis tools but not a great deal of guidance on method. This book is exactly what I was looking for to fill that gap.

[Moahmmed · Rating 4 out of 5]

It's a good book for anyone who wants to know more about data science and data science analysis
In this book, Roger D.peng showed the entire process of data science analysis :
1-stating a question
2-EDA (Exploratory data analysis)
3-Using Models & Expectations
4-inference and prediction
5- Interpreting Your Results
6-compunctions

[Ray · Rating 2 out of 5]

This is an introductory book on how to think analytically and some of the terminology that goes along with it. It's good for learning how to speak data science and data analysis, but it won't get much further than that. It's helpful for touching up on your ability to think analytically, refresh on terminology, and hone presentation skills.

[Alejandro Donaire Salvador · Rating 3 out of 5]

It is a useful and interesting book overall, well-structured, easy to get through and visually pleasant. Most importantly, I learnt a few new things and ideas. However, sometimes I found the content to be repetitive or obvious, and it felt more like a review of the basics of Data Science. Besides, there were a few typos. I recommend it to beginners, but not so much for more experienced people.

[Klaus · Rating 2 out of 5]

Basically a short research methodology book utilizing data science.

If you're already familiar with the former topic, there's quite a lot of filler and it's a short book in the first place. If you're not, it'd be a decent intro to the topic.

[Nora · Rating 4 out of 5]

There is a lot of helpful information pertaining to all steps of performing a data analysis. The information is well written and clear with lots of examples to help with understanding. Will be a good source for reference.

[Ayobami · Rating 5 out of 5]

Great book for those new to data science. The illustrations and examples were apt.

[Mustafa Hajmohammed · Rating 5 out of 5]

Gentle introduction for those who are searching for more knowledge on data science in general and data mining in particular.
Not recommended for experienced data scientists, it is too basic.

[Ferhat Elmas · Rating 3 out of 5]

High-level useful bits.

[Ioannis · Rating 4 out of 5]

An informative book on the process of data analysis from start to end.

[Léah · Rating 4 out of 5]

Really clear with helpful examples. Good introduction for approaching analyses from an artsy perspective in simple language

[Mahmoud Galal · Rating 4 out of 5]

It's simple and very concis.

[harish · Rating 4 out of 5]

Simple and concise book for anyone who is just starting their journey into Data science field.

[Juan Jose · Rating 3 out of 5]

Interesting and a great first approach to data science

[Afifah Luqman · Rating 4 out of 5]

Great read on basic data analytics. Walk through of thought process. very useful

[Lukas · Rating 3 out of 5]

Quite interesting framework, although too elementary for experienced data scientists

[Caitlin · Rating 2 out of 5]

The first few chapters about determining the type of question being analyzed were well throughout and informative but the rest was lacking.

[Joren Van Roy · Rating 3 out of 5]

Not so practical as I hoped it to be.

[Kristine · Rating 5 out of 5]

This is an excellent book that breaks down the steps of data analysis. Many of these steps are carried out intuitively by data analysts, and it was enlightening to have them identified and put into context of the process. As a data analyst, I enjoyed this book. I think it would be a great read for people who work with analysts but don't have a clear view of what the work entails; I believe it would improve their appreciation of the creativity and complex processes carried out by the analysts with whom they collaborate.

[Ray · Rating 4 out of 5]

I read this in South Africa when it was recommended reading during the first week of a data science course. While it's hard to know for certain, I believe that the perspective I gained from it contributed towards my success in the course. I learned the most about all the non-science parts of a data scientist's job, and this was valuable to me later as I began to accumulate some of my earliest work experience in a data-science-like role.

[Fermin Quant · Rating 4 out of 5]

Great book with awesome examples. There are quite a few errors in grammar and spelling, but they do not subtract from the value of the knowledge. Very useful frameworks to understand and apply to data analyses.
I did feel the last third part was a little rushed, and like it is somewhat incomplete, but overall the content was really good and useful.

[Yahia El gamal · Rating 3 out of 5]

This is a light data science book. I have mixed feelings towards it. Sometimes, Prof. Roger writes assuming some prior knowledge and experience (e.g. he wrote some parts freely assuming some knowledge about distributions, significance, ..) and those are the parts I really liked in the book. On the other hand, prof. Roger sometimes writes as if the audience are total newbies with no previous experience.

Leaving the above note aside, the abstraction Prof. Roger introduces about the process of data science / data analysis will probably be agreed upon by everyone. It probably won't add much to experienced people (as it is probably a second-nature) except when they start introducing data science and data analysis to others, and for this reason, i recommend this book. The abstractions and the examples will make introducing someone to the field a much organized process.

The book is highly accessible, enjoyable, and short. You won't invest much in reading it (time-wise) and you will get decent output from it.

Disclaimer: I read a pre-released version from leanpub