Practical Time Series Analysis: Prediction with Statistics and Machine Learning

This book was just released and I'm only a few chapters into it, but I can already attest that it's a fantastic treatment of the topic. It reads like the book I would have written (or perhaps would have liked to have written) if I were to pen a volume on time series analysis. It's a long overdue contribution to the subject, and I'm happy to see it executed so well. It's very well organized and written in an approachable style, yet does not shy away from the details. It's clear that the author has a great deal of practical, applied experience and has "been there, done that." For example, the chapter on data cleaning impressed me not only because she has the kind of background that impresses on one the importance of clean data, but also because the discussion mirrored many of the hard-earned lessons that I've wrestled with over the years. If you're interested in time series analysis as more than a purely academic exercise, you would be well placed to have this book on your bookshelf (or in your Kindle library, as the case may be). I highly recommend it.

I'm about halfway through this book but it is a very easy read and great introduction into time series. A lot of books assume prior knowledge but this one feels like it takes many books and puts them together. It feels like three main segments: an intro to what time series is and the data processing steps, then the introduction of various models and forecasting, and the last segment being the application into machine learning. This is a very great read and I'm happy I purchased it. I would also like to say that those that complain about the book being written in 2 languages. This is not the book for you if you find it a burden to switch between R and Python. Do not use this book as a beginners guide to programming. That's not its purpose. In my opinion the book isn't hindered by using both but is better because you understand the significance of the libraries being widely available. Each language has its benefits and cons. And in my own career I've switched between both.

Difficult to have an higher opinion of the book. The global feeling is that the content has not been reviewed with the care needed for such topics. The author blended very superficial mathematical explanation for several approaches, with more detailed maths for others. Same comment for the R vs. Python code examples. And I found also the tendency to illustrate the content with examples and codes producing very very low performing models. It is often a tendency now to have people including into their books codes that produce something that ... does not work, or works with very low performance. And the excuse is always the same : "we should spend more time on the days or the model....". If they know that their chosen examples do not work good, they should have not included them in their book, or do a better job. So, at the end, I read this book cover-to-cover, and I really do not understand how si many people have rated it so high.

Ms. Nielsen is an excellent writer and this book is a (much-needed) introduction to the science of time-series analysis. Ms. Nielsen presents the concepts as well as the tools and techniques and is presented in a practical, problem-solving manner. I highly recommend this book. One of the best O'Reilly series books I've purchased.

Highly needed book which focuses on core time series analysis along with its applications in different domains. Covers a huge range from time series data wrangling, feature selection to building statistical models. Suitable for people who have some knowledge of python, machine learning, time series and neural networks.

This is a good book on time series data analysis for entry-level students if you are comfortable with R and Python, which are too slow in performance in general. It covers a lot of subjects to which time series data analysis is applicable. However, I did not find it very helpful for me as I was looking for more in-depth coverage of forecasting market moves, which are events of time series nature. If you are looking for something of similar interest, check out Forecasting and Timing Markets: A Quantitative Approach , which is dedicated to the so-called mechanical trading mentioned on page 9 of this book.

Easy read for someone with a ML background that gets straight to the point of how to approach time series analysis! Covers multiple approaches and common time series specific pitfalls.

An excellent overview of time series. It is clear that the author put a lot of effort into this book. If you are new to the subject and already have some coding background this is a good starting point.

The author mentions some important topics only in passing, such as the I in ARIMA. Not a single equation defines an ARIMA model in the section on ARIMA models. Looks like the author wrote it in one go, without further editing.

Está bien explicado

This book has quite a few chapters using R coding only. Some other chapters use python. If you only know python, then you do miss few chapters. I hope every chapter comes with python codes.

It is not at all a book! It is a set of notes and jargon. She (Author) did not even explain a single topic properly. In many of the chapters, it is written like this 'This is out of scope of the book'. Quite surprising. Then what the book is for ? For example, AR(1) and AR(p) process not explained clearly. What's the role for order p ? It is very important is time series. She has just used some complicated language to explain a simple topic. Mathematical expressions are loosely made in latex editor and plugged in randomly which looks very odd. Either the author did not have any experience of the subject. Possibility is there that she has collected notes from here & there and just compiled as a book. The reason could be that she has a poor writing skills. It is quite unfortunate that proposals of many good prospective authors are rejected by Oreilly and after that they publish this type of book

I’m writing this review because I was so excited this book existed right when I needed to ramp up on time series methods but the bad reviews meant I couldn’t say nothing. Yes the book alternates between Python and R code. The code snippets are small and read like psuedocode so knowing one should make it easy enough to port it to the other. This is not a beginner book. It’s giving a history of time series analysis and bringing you up to speed with modern techniques. I felt it was the right book for me as it explained the intuitions behind the maths better than I had previously been exposed to. I even implemented my first ARIMA model to model airconditioner usage in my house with a RaspberryPi and a DHT22 sensor. But I was also able to use the strategies in Ch11 in a project at work to evaluate our forecasting and measure error. As it says on the tin, it was practical timeseries analysis. There are quite a few other timeseries text books listed in this book if you want other perspectives or want to go deeper academically. Get a kindle sample and read the first few chapters for that list of books. I would wholeheartedly recommend this book as a reference and a bridge for those wanting to add timeseries to an existing tool belt of analytical skills. It is not a beginners book though. I have the kindle version but I am buying the paperback to keep on my desk as a reference for teaching interns.