Data Science

The MIT Press Essential Knowledge

by John D. Kelleher, Brendan Tierney

A concise introduction to the emerging field of data science, explaining its evolution, relation to machine learning, current uses, data infrastructure issues, and ethical challenges.

The goal of data science is to improve decision making through the analysis of data. Today data science determines the ads we see online, the books and movies that are recommended to us online, which emails are filtered into our spam folders, and even how much we pay for health insurance. This volume in the MIT Press Essential Knowledge series offers a concise introduction to the emerging field of data science, explaining its evolution, current uses, data infrastructure issues, and ethical challenges.

It has never been easier for organizations to gather, store, and process data. Use of data science is driven by the rise of big data and social media, the development of high-performance computing, and the emergence of such powerful methods for data analysis and modeling as deep learning. Data science encompasses a set of principles, problem definitions, algorithms, and processes for extracting non-obvious and useful patterns from large datasets. It is closely related to the fields of data mining and machine learning, but broader in scope. This book offers a brief history of the field, introduces fundamental data concepts, and describes the stages in a data science project. It considers data infrastructure and the challenges posed by integrating data from multiple sources, introduces the basics of machine learning, and discusses how to link machine learning expertise with real-world problems. The book also reviews ethical and legal issues, developments in data regulation, and computational approaches to preserving privacy. Finally, it considers the future impact of data science and offers principles for success in data science projects.

Genres: Nonfiction, Science, Technology, Computer Science, Computers, Artificial Intelligence, Business

280 pages, Paperback

Published April 13, 2018

About the author

John D. Kelleher is a Professor of Computer Science and the Academic Leader of the Information, Communication, and Entertainment Research Institute at the Dublin Institute of Technology. He is the coauthor of Fundamentals of Machine Learning for Predictive Data Analytics (MIT Press).

Reviews

[☘Misericordia☘ ⚡ϟ⚡⛈⚡☁ ❇️❤❣ · Rating 4 out of 5]

Not technical enough. No examples of code or even pseudocode. Basically, this is an elementary intro to DS. A very accessible one, mind it, which is a rare-ish attribute. Still, the book definitely could have done with more depth.

[Randy · Rating 4 out of 5]

There are many perspectives on what Data Science actually is. This book helped me prepare a lecture on the topic by covering a more-or-less standard view that data science is the combination of many related math/stats/machine-learning/programming fields combined with domain expertise. The footnotes and references were especially helpful. One of the best was: Battle of the Data Science Venn Diagrams. by David Taylor on KDnuggets, October 2016.

This is not a book that I would recommend to anyone that knows much (anything?) about data science. It is a fine introduction for people coming from outside of the field. There are too many practical data science books to mention even the really good ones, but one of the best I've taught from is: R for Data Science: Import, Tidy, Transform, Visualize, and Model Data which is simply outstanding (there is also a free online version: at the author's website). I recently picked up another interesting book The Data Science Handbook, which uses Python goes into more detail on machine learning methods.

If you dig a little you might think that you have to choose between Python and R for this sort of work, but this is no longer required. Historically R has often been paired with other languages when data wrangling, text munging or database interfaces were better done in another language (a lot of people learned Perl to do this in bioinformatics). Ruby is also good for glue code in data science projects, but Python has taken over as the most popular for many high-level tasks (especially in ML thanks to Google and the TensorFlow library). In my humble opinion R remains stronger for pure stats work. The real point is that you no longer have to choose. The two languages play extremely well together. So pick one (or both), there is no shortage of books on both of them.

[Akbar Madan · Rating 4 out of 5]

ركيزة أساسية في عصر الذكاء الاصطناعي
في عصرنا الحالي، أصبحت البيانات بمثابة الوقود الذي يغذي محركات التقدم والابتكار. فالحكمة الإنسانية، التي تتجلى في اتخاذ القرارات الرشيدة وحل المشكلات المعقدة، تعتمد بشكل أساسي على المعرفة المستمدة من البيانات المنظمة والمحللة.
لم يعد علم البيانات مجرد تخصص أكاديمي، بل أصبح ضرورة حتمية في ظل التطور المتسارع للذكاء الاصطناعي. هذا الذكاء، الذي يعيد تشكيل سلوكياتنا وقراراتنا، يعتمد بشكل كبير على البيانات الضخمة وتحليلها الدقيق.
إن جمع البيانات وتحليلها يفتح آفاقًا واسعة لتحقيق التقدم في مختلف المجالات، سواء كانت اقتصادية، اجتماعية، ديموغرافية أو علمية. فمن خلال تحليل البيانات، يمكننا استخلاص الأنماط والاتجاهات التي تساعدنا على فهم الواقع بشكل أفضل واتخاذ قرارات أكثر فعالية.
يلعب الذكاء الاصطناعي دورًا محوريًا في تعزيز قدرتنا على التعامل مع البيانات الضخمة. فهو لا يساعد فقط في جمع البيانات وتخزينها، بل يساهم أيضًا في تحليلها بطرق متقدمة، مما ينتج عنه رؤى قيمة تساعد في رسم خرائط رقمية شاملة لمختلف الأنشطة الإنسانية.
تهدف هذه الخرائط الرقمية إلى تحقيق الانتظام المجتمعي المستدام، وذلك من خلال توفير معلومات دقيقة ومحدثة لصناع القرار في مختلف القطاعات. فمن خلال فهمنا العميق للبيانات، يمكننا تطوير حلول مبتكرة للتحديات التي تواجه مجتمعاتنا، وتحقيق التنمية المستدامة التي تنعم بالسلام والعدالة.

[Uriel Vidal · Rating 5 out of 5]

Libros, Amazon y yo

Siempre he tenido esa sensación que un libro te escoge, es decir que un día cualquier vas caminando por algún lugar y ves un libro que en tu vida has visto o te ha comentado, pero algo te llama la atención de él que decides recogerlo y comprarlo. Esto me pasa muy seguido en las librerías, ya que hay más libros y por lo tanto hay una probabilidad mayor de que un libro te escoja.
Este libro no vino de una librería y ni de verlo en alguna caminata, sino vino de una recomendación de Amazon. Con cualquier recomendación de me hace Amazon generalmente lo que hago es reviso las críticas que han hecho en su propio sitio y en goodreads (que creo que igual le pertenece a Amazon), si es que el libro resulta interesante generó una revisión histórica de precios para verificar que sea un buen momento para comprarlo o en su caso le genero un seguimiento para poder comprar cuando este en un precio mínimo.
Con este libro hice este mismo proceso, pero lo que me ocurrió fue algo curioso y es que el libro tiene su versión en inglés, lo que es comprensible por el tema, y además de ambas versiones no se llevan demasiado tiempo de una a otra como sucede con otros libros, sin embargo lo curioso de esto es que los dos tenían un precio similar (15 pesos de diferencia cuando hice la consulta, pero menos de 2 cuando hice la compra), pero uno era vendido por Amazon EUA y el otro por Amazon México, y el libro en español era editado en EUA, además obviamente de tratar el mismo tema y ser de la misma editorial.
Se me hace curioso esto ya que cómo es que Amazon me haya recomendado un libro en español, cuando existe la versión en inglés que está mejor calificada en su sitio, sea un primera edición, editada en EUA y además vendida con el sitio de Amazon EUA. Mi primer pensamiento fue que era por el precio, pero ¿realmente valen esos 15 pesos extras toda la logística que implica traer un libro desde EUA? ¿O es porque hablo español y claro está, sería mejor algo en mi idioma materno?
En fin, estuve con eso en mi mente y decidiendo que versión comprar: la versión en inglés con el atractivo de poder tener una referencia a la mano con conceptos básicos y sus acrónimos, que se usan mucho en este tipo de libros y que muchas veces se usa ese acrónimo en vez del concepto o comprar la versión en español que me daría igual esos conceptos a la mano, pero con la posible pérdida de información de esos acrónimos que son tan útiles cuando regresas a un libro por algo de información y con temor de que la traducción no sea tan acorde a la idea que presenta el autor... Me decidí por el libro en español.

La ciencia de datos como sistema

Este libro trata de la ciencia de datos como un todo, generalmente cuando nos comentan acerca de la ciencia de datos generalmente se nos viene a la mente cosas como algoritmos de aprendizaje, redes neuronales, aprendizaje profundo, etc, y claro está toda la matemática asociada a ella. Claro está que este libro trata de eso, sin embargo, no es lo único de lo que trata ya que además revisa los procesos que están antes de que inclusive pasemos a la recolección de datos y las implicaciones éticas y morales que contrae tener un proyecto de ciencia de datos.

Para muchas personas el hecho de tener un libo que no hable con ecuaciones, de matemática o de algoritmos en este campo hace que sea considerado un libro de dudosa calidad, sin embargo, está parte de la ciencia de datos es la que menos trabajo lleva dentro de un proyecto de ciencia de datos y es que muchos de estos algoritmos ya se encuentran hechos y como se dice: "no se debe reinventar la rueda cada vez". Toda esta parte que está detrás de los proyectos de ciencia de datos, de dónde se extraerán los datos, los tipos de datos que se necesitan, la limpieza y carga de los mismos y si realmente podemos usarlos para los fines que perseguimos son incluso más importantes para los proyectos de ciencia de datos, ya que estas actividades son cruciales para que tengamos un proyecto de ciencia de datos exitoso.

El libro hace una excelente traducción de los conceptos tratados en la ciencia de datos y aunque tiene ciertos errores en la traducción en el sentido de que quisieron respetar lo más posible el texto original que no hace mucho sentido en español. Igualmente con los acrónimos se hace un excelente trabajo ya que se respeta el acónimo en inglés, pero se coloca el significado el español, lo que hace que sea sencillo el que este libro sea una referencia futura.


La luz en la ciencia de datos

Existe un viejo chiste concerniente a la ciencia de datos: “El Big data es como el sexo en la adolescencia: todo el mundo habla de ello, nadie sabe realmente cómo hacerlo, todos piensan que los demás lo están haciendo, así que todos dicen que también lo hacen”. Me parece que este chiste se puede extrapolar a la ciencia de datos, ya que los que estamos interesado en ella, generalmente tenemos una barrera de entrada alta, puesto que todos la mayoría de los cursos y videos hablan de ella como una arte mística y obscura que solo los matemáticos y algunos magos pueden entender.

Al leer este libro me di cuenta que muchas cosas de ciencia de datos son cosas que ya sabía previamente, como los algoritmos de aprendizaje o las redes neuronales, inclusive cuando en algún trabajo una de las actividades clave revisar a profundidad si es posible usar los datos con lo que se iban a trabajar.


Un toque no todos tienen

Un libro que hace da referencias a más libros se me hace un gran libro y es que indica que los autores indiquen un camino por el que podemos continuar con el aprendizaje. Este libro no solo tiene referencias dentro del texto, sino que tiene una sección completa de lecturas adicionales, además como un buen texto académico, tiene todas las referencias en usan para poder ir con más profundidad en los temas que mas nos interesen.


¿Es para mi?

Este libro va dirigido a personas con que tienen el interés por adentrarse en la ciencia de datos, esas personas que han hecho quizá un par de cursos online, pero se quedaron con dudas acerca de que si entendieron todo y que existe más allá, el libro maneja muy pocas formulas, pero si es necesario tener algo de contexto de programación, como el saber qué es una instancia o una base de datos. Es un libro de entrada, de consulta y referencia, además de que da una visión global de lo que es y de lo que no es ciencia de datos.

Pensamientos finales

Para finalizar la historia inicial, después de leer el libro pude darme cuenta todo lo que Amazon conoce de mi, como es que yo con mis gustos y compras que he hecho en la plataforma, pudo recomendarme un libro del cuál no tenia idea que existía, de una edición que no hubiese comprado y con un precio más alto del que hubiera comprado, da algo de miedo saber lo mucho que saben de uno, pero también es interesante saber cómo esa recomendación puso en mis manos un libro que disfruté bastante.

[Justin · Rating 4 out of 5]

I think a more clear title to this book would be "Introduction to Data Science". While one may be able to infer that, it was definitely written for an audience that knows very little about what Data Science is and is looking to have a general overview. For that, I felt the book did a very good job. It covered a lot of the basic topics, but none of them very in depth (no math! haha). It also had good discussion around the business and ethics sides of Data Science, as well as its ongoing integration into society.

It was probably a little more basic and overview style than I was really looking for given my current knowledge, but a good resource for anyone thinking about moving into Data Science or just wanting to know more about it.

[Adrian Woods · Rating 4 out of 5]

Great introduction. Not too technical.

[RAD · Rating 5 out of 5]

Appropriately Titled

Data Science is a sterling examplar of the excellent MIT Press Essential Knowledge Series (itself an analog of the also excellent Very Short Introductions series from Oxford University Press). Both series present concise overviews of a particular topics from a leader in the field; and coming as they do from academic presses, tend to have both Notes (inexplicably, endnotes rather than footnotes--when will this insanity end?) and Further Reading sections (Data Science adds a Glossary which some readers may find helpful).

The book is structured as seven brief chapters:
1. What Is Data Science?
2. What Are Data, and What Is a Data Set?
3. A Data Science Ecosystem
4. Machine Learning 101
5. Standard Data Science Tasks
6. Privacy and Ethics
7. Future Trends and Principles of Success

The first and last chapters were, for me, the most helpful. In the real world, "data science" is ill-defined. In theory, it spans many disciplines: computer science and programming, machine learning, statistics, mathematics, and logic. It also includes subjective attributes including creativity, communication, critical thinking, and domain expertise. In practice, the definition is more straightforward: an ability to code, along with some statistical knowledge.

Chapter 1 begins with a definitional answer: "Data science encompasses a set of principles, problem definitions, algorithms, and processes for extracting non-obvious and useful patterns from large data sets" (p. 1). It provides a historical overview ("dating back to the 1990s", p. 6) and highlights areas of use (government, sports, and "nearly all parts of modern societies", p. 24). It notes data science's definitional confusion by pointing to a public lecture in 1997 which proposed that the field of statistics should be renamed "data science" (p. 17). And there is an excellent illustration (Figure 1, p. 20) that depicts many of the characteristics that a data scientist should have.

(With 20 years of experience in the field, it has been my observation that the practical definition of a data scientist is far less robust, focused only on the coding and statistics attributes. Think of a data scientist as a chef. A chef combines myriad ingredients to produce something new (and useful) to others. Yet few, if any, chefs are organic chemists--this humorous example with Julia Child notwithstanding. The vast majority of today's employed data scientists are organic chemists; the world of data needs more chefs).

For current or aspiring data science practitioners, Chapter 7 is worth the price of the book. In a subsection titled "Data Science Project Principles: Why Projects Succeed or Fail" (pp. 225-ff.), the authors--who can only be speaking from hands-on experience--outline the necessities of any data science project. Ironically, for such a statistically focused undertaking, success is not just a function of the statistical or mathematical goodness of a particular model; rather, it incorporates several exogenous factors, including management support, explainability to a constituency, teamwork and input among stakeholders, business integration, and other subjective factors.

Like other titles in the series, Data Science should be helpful to a broad range of users. While the specialist or practitioner won't learn any technical tidbits, they should have their existing views reinforced (or new ones illuminated); and for the layperson, it is an excellent overview.

[Linkers · Rating 3 out of 5]

Solid introduction into the different areas of Data Science.

[kayley anderson · Rating 4 out of 5]

really great overview of data science and issues that can arise, definitely meant as an introduction but still had a good amount of detail and explanation

[Pablo · Rating 4 out of 5]

Great intro from non experts.

[Janne Heinonen · Rating 4 out of 5]

Vallan mainio tietopaketti, ainakin tämmöiselle tietojärjestelmähenkilölle. Terra Cognita on tehnyt taas hyvää työtä, ei olisi tullut varmastikaan alkuperäiskielellä luetuksi.

[Deana · Rating 4 out of 5]

for a mostly nontechnical audience, but a good summation of the basic practices, principles and objectives regardless

[Grace Tierney · Rating 5 out of 5]

It has been 18 years since I worked in databases and computing so this slim volume enticed me to update my knowledge of the growing field of Data Science. The series of slender, beautifully produced books (lovely paper, clean font etc) cover all manner of topics at an introductory level and are a handy way into a new field. This was not a disappointment.

I was delighted to find I recalled the basics re data warehousing and very large databases, but also to get the fundamentals of neural nets, and newer tech underlying the buzzword area of data science. The chapter on ethics was particularly fascinating and the book is filled with case studies which illustrate their points. If you think data science doesn't affect your life then think again. It underpins customer loyalty schemes, companies predicting which customers will change to a new provider (and how to prevent that), policing, medical advances, traffic flow management and much more.

You won't find code here, it's an overview and introduction to the key topics, but even for nontech readers it should provide a good answer to What is Data Science and Why should I Care?

Disclosure - I know one of the authors, but this is an honest review

[Edward Soon · Rating 4 out of 5]

This book makes me realize how the statistical subject can be taught in a different way. (way better) i had never been taught on the application for the analytics models in real life.
This book is definitely a good 101 handbook to begin with. This book uses most of the layman terms and sensible examples to explain the essential fundamental jargon, ideas, and concepts required to understand data science.
My favourite part is the privacy and ethics. It gives me more in depth insights about the dark side of data science (when everyone sees data driven technology is appealing).

[Matt Heavner · Rating 2 out of 5]

OK, but both too broad and too technical. If you are familiar with data science / machine learning (ML) at all, this book is too basic. But if you aren't familiar at all with data science/ML, this book is too technical. With a good discussion of ethics, bias, and limitations of ML; this book still tried to cover too much ML space with too much detail and also not enough. Overall, I'd say a good attempt, but it missed the mark.
And three really horrible typos in the book - editor?

[Narinder · Rating 5 out of 5]

A primer for amateurs as well for experts.

[Chris · Rating 3 out of 5]

This is a pretty dry and not-very-well written intro to data science. It does a reasonable job of explaining most of the definitions, although it will frequently throw in terms it hasn't yet defined at that point. It's completely stripped of explorative narrative structure, with the occasional deterministic example thrown in here and there.

It may also repeat introductions to ideas, like it did with Hadoop, where it introduces it twice, only a few paragraphs apart. It's also strangely disproportionate in its introductions. For example, it discusses OLAP and SQL as if they were two sides of a boring slice of bread, and then refers to in-memory ML with such strong golden-egg rhetoric that you wonder if the author has an in-memory ML start-up.

The chapters were also too long. Although they utilised sub-sections, the chapters could have been broken up further for ease of mental processing. It could also have used chapter summaries and/or comprehension questions to ensure you've understood the key points. This was really needed as is chapter was a dense bush of not-well organised collections of concepts and definitions.

In short, this needed a more professional/competent writing and editing team before publication.

What did I try to gleam from this hunk of words?

Data
• Data science is seeking nonobvious and useful findings (actionable insights) from data through pattern extraction.
• Common uses: customer segmentation (clustering); association rule mining; anomaly/outlier detection; and classification/prediction.
• 3Vs of big data: Volume of data, Variety of data, and Velocity of processing.
• Data: numerical, nominal and ordinal; structured and unstructured; and raw data: captured and exhaust data.
• DIKW pyramid: Data (abstract elements) preceeds information (linked/structured data), which preceeds knowledge (organised information), which preceeds wisdom (applied knowledge).
• CRISP-DM: 6-stage semi-structured cycle with data at its centre: business understanding, data understanding, data preparation, modelling, evaluation and deployment.
• Data science ecosystem: data source/ capture applications; data storage (structured or unstructured db); and data processing applications (eg. analytics, back-end, business processes).
• OLAP: Online Analytical Processing. A processing system used in data warehouses to generate summaries, and aggregate, slice, dice and pivot the data. Often uses a data cube: fast but rigid fixed dimensional data processing. SQL is often used as a less restrictive alternative.
• Hadoop: big data distributed storage and analytics processing. Uses MapReduce based on the split-apply-combine strategy, where each split data chunk is stored on different nodes in a parallel processing cluster.
• In-database ML: Avoids moving or formatting the data for ML processing, increasing processing speed and reducing errors.
• Alternatives to Hadoop for big-data include: Storm, Spark and Flint. While Hadoop uses data-set batch processing, better suiting summaries of static data, Storm uses individual-item stream processing, better suiting continuous classification tasks. Spark and Flink are hybrids, using both batch and stream processing.

Machine Learning
• Linear regression: when input attributes are fixed, a regression function's parameters can be estimated using individually weighted averages for each point as they best fit a linear regression function. Often uses least squares for error estimation.
• Neural networks: each neuron takes a sum of the weighted inputs and outputs them to an activation function (logistic function or tanh).
• Training neural networks uses backpropagation, where the weight-update rule adjusts the weights and error term in each previous neuron.
• Recurrent neural networks (RNNs) use feedback loops to remember the context of the previous inputs, often used in NLP.
• Convolutional neural networks (CNNs): each set of neurons share their weights to identify a distinct visual feature. The outputs of each layer serve as inputs to the subsequent layer.
• Decision trees: best with nominal and ordinal data; uses classification rules based on data attributes to determine labels. Uses a recursive approach, adding nodes as needed, from root to leaf until all labels on the same branch agree.
• Concept drift: contrary to the stationarity assumption, behaviours change over time. Models that were previously trained and accurate need to be periodically retrained to adapt.

Data Science Tasks
• Clustering: sorting data into subgroups, such as customer groups. Eg. k-means clustering.
• Anomaly/outlier detection: detection rules for anomylous events (eg. fraud) are checked against data in the data warehouse. Often written in SQL (such as MATCH_RECOGNIZE). Prediction models can also be trained for concurrent detection. Eg. Support-Vector Machines (SVMs) or One-Class Classifier algorithms.
• Association-rule mining: unsupervised learning for detecting groups of items that frequently co-occur (eg. Recommendation algorithms and cross-selling systems). Can be used to explore purchase behaviours through correlations of attributes. Eg. Apriori algorithm.
• Prediction (classification and regression): exploring the likelihood of future action (such as customer churn or price prediction). Eg. Customer-propensity models.

[russell · Rating 4 out of 5]

I’m a huge fan of series like The MIT Press Essential Knowledge Series and Oxford’s A Very Short Introduction series that provide a detailed but still general birds’ eye view of interesting topics. This book, like many others in those series, help complicated ideas stick by bundling them with a narrative, broad view of a topic. Also they are great because they are shortish.

Throughout Data Science you can tell Kelleher and Tierney have applied data science experience in a business setting, rather than just in an academic one. Having both perspectives made for a more engaging read without sacrificing on technical depth.

I wasn’t expecting a portion on ethics, but I’m glad they included it to promote best practices. Lord knows business people need to be reminded of moral boundaries, and I suspect a nudge of a few degrees (thank you Thaler) in the right direction early in DS practitioners’ careers can make for large gains for society later on.

[eri b.❀ · Rating 4 out of 5]

The goal of data science is to improve decision making by basing decisions on insights extracted from large data sets. As a field of activity, data science encompasses a set of principles, problem definitions, algorithms and processes for extracting nonobvious and useful patterns from large data sets.

What an excellent introduction to data science! I really, really couldn't have asked for more (except maybe some practical examples pretty please?). Even not knowing much about data science (and maybe because of that) this was a light lecture, full of good info and the best place to start learning for real. It gives you the introduction, explanation of some algorithtms and a long, good talk on Data Science's ethics. A very helpful lecture if you are interested in Data Science, or if you just want to understand the concept.

[Piritta · Rating 2 out of 5]

Luin suomenkielisen käännöksen, ja sen heikkous vei täysin huomioni itse kirjalta, joka on ilmeisen tekninen tällaiselle asiaan heikosti perehtyneelle, mutta luultavasti liian yksinkertaisesti kirjoitettu ollakseen hyödyllinen ammattilaiselle. Käännöksessä oli kritiikittömästi valittu "datan" ainoaksi käännökseksi "aineisto", ja se meni monessa kohdassa pahasti pieleen. Erityisesti luku gdpr:stä oli hirveää luettavaa, koska "personal data" on vakiintuneesti suomeksi "henkilötieto", joka on selkeästi sekä käsitteellisesti että juridisesti eri asia kuin kirjassa hoettu "henkilökohtainen aineisto". Käännös on muutenkin raakakäännösmäinen, interferenssiä on paljon ja sisältö jää sen vuoksi monessa kohdassa aukeamatta lukijalle. Tällainen käännös tekee pelkkää hallaa käännösalalle.

[Cai Tianji · Rating 5 out of 5]

Must-read for anyone who wishes to enter data science

Well-written and easy-to-understand, this book gives a new-comer like me a conceptual framework to think about problems in data science. It helps me to understand what the field really is and what the workflow of a data science project looks like. Particularly interesting is the chapter on data ethics and regulation. I think it is an area that is often overlooked by technical textbook, but should really be emphasized to readers who might someday become a data practitioner. Overall, it’s a very good book and worths your effort to delve into.

[Matias · Rating 4 out of 5]

How can data science predict the future?

I stumbled upon this book from my Kindle suggestions page and I found it really useful on my studies. MIT press essential knowledge series are a good way to get a bigger picture on a subject. Data science book covers very well what is data science, what are the main methods and how is a data science project organized. As a Computer science student I would loved to know more about the algorithms, but I understand that the main goal on this book was to tell about data science to whomever wants to know about data science. If you want to know how can data science find new information from various sources that human cant check this out.

[path · Rating 4 out of 5]

A concise and accessible introduction to data science practices. The authors do a good job of explaining modern, situated interests that are driving the explosion of data science. It is in that context of commercial, medical, and civic motivations that the various data collection, analysis, and modeling techniques are described. All of them solve particular kinds of problems, and understanding the problems helped me build an intuitive sense of how the science works. The authors also provided a nice chapter on the problems created by data science, particularly about privacy and ethics. The issues are presented clearly and with an appreciation for their gravity.

[Johannes · Rating 3 out of 5]

Start: 3/5
Middle: 5/5
End: 2/5

As an introduction to data science, it work's quite well for any type of person. I think the start is a bit to 'corporate' for me (as a developer), but should be fine for a project manager etc. The middle part, the actual machine learning was really well. Then it ends with (data) ethics and project management, which frankly, was dry. Again; for a project manager It would most likely be great, but for an ethics nerd, way to dry.

Overall a quite good introduction book that I'm not afraid to recommend to anyone.

[An Te · Rating 4 out of 5]

A good introduction to data science including the various algorithms types, neural networks (from neutrons) and detailing the importance of privacy in the use of data. A helpful list in the final chapter gives good advice on how to do good data science. Reading this will give you a foundation from which to launch into more substantial texts.

I'd add that this is mostly from the US perspective, and so may not reflect the dat science (and ethics) of your nation.

[Horacio Garza · Rating 4 out of 5]

Buen libro para las personas que van incursionando en el mundo de Data Science o para aquel uppermanagement que busca implementarlo en su organización. Cualquier persona que busque incursionar en la parte de desarrollo de modelos este no debe de ser su único libro, ya que NO es un libro técnico pero mas de conceptos, debe de incluir con un montón de lobros de estadística que ayuden a entender todo el data science de forma eficiente.

[Christopher · Rating 5 out of 5]

Really fascinating and easily digestible introduction to data science. The most difficult chapter in this book is about artificial intelligence, but it's still understandable even without advanced knowledge of computer science or statistics. Everyone who's working with databases or in business analysis should read this.

[Joseph · Rating 5 out of 5]

A great introduction to the field of data science. Certainly keeps the terminology quite rudimentary, but displays the fundamental understanding one needs when learning more about cleaning/preparation, ML, and the ethics behind large scale uses of data. Highly recommend to anyone wanting to peek into the hyped field of data science

[Anthony O'Connor · Rating 4 out of 5]

Solid

A solid well written introduction to the emerging field of data science. Math statistics and big data. Machine learning. Neural networks. Deep learning. Privacy regulations. All outlined and placed in context. A refreshing and sober level of order and clarity without the hype.

[Peter Aronson · Rating 4 out of 5]

A solid introduction to the subject, that does not shy away from practical and ethical issues. It is perhaps a bit more positive about the uses of Data Science than I would be, but that is not surprising, since you wouldn't generally write a book like this unless you felt that Data Science was overall a positive thing.