返回到 Practical Time Series Analysis

星

1,469 個評分

•

418 條評論

Welcome to Practical Time Series Analysis!
Many of us are "accidental" data analysts. We trained in the sciences, business, or engineering and then found ourselves confronted with data for which we have no formal analytic training. This course is designed for people with some technical competencies who would like more than a "cookbook" approach, but who still need to concentrate on the routine sorts of presentation and analysis that deepen the understanding of our professional topics.
In practical Time Series Analysis we look at data sets that represent sequential information, such as stock prices, annual rainfall, sunspot activity, the price of agricultural products, and more. We look at several mathematical models that might be used to describe the processes which generate these types of data. We also look at graphical representations that provide insights into our data. Finally, we also learn how to make forecasts that say intelligent things about what we might expect in the future.
Please take a few minutes to explore the course site. You will find video lectures with supporting written materials as well as quizzes to help emphasize important points. The language for the course is R, a free implementation of the S language. It is a professional environment and fairly easy to learn.
You can discuss material from the course with your fellow learners. Please take a moment to introduce yourself!
Time Series Analysis can take effort to learn- we have tried to present those ideas that are "mission critical" in a way where you understand enough of the math to fell satisfied while also being immediately productive. We hope you enjoy the class!...

SS

2021年4月6日

It is a very good course which builds on the basics of time series and also covers more advanced topics like SARIMA. The course contains ample examples which helped me better understand the material.

SA

2020年1月23日

Excelente, uno de los mejores cursos que he tomado. Lo más importante es que se practica muy seguido y hay examenes durante los vídeos. Si hay un nivel más avanzado de este tema, seguro que lo tomo.

篩選依據：

創建者 Alireza P

•2020年7月22日

Perfect

創建者 ِِِAli A A

•2020年7月16日

perfect

創建者 Douglas B P

•2018年9月2日

Great!!

創建者 Cathy D

•2018年5月22日

useful!

創建者 Alla E G

•2020年1月18日

Thanks

創建者 Mehrpouran, S (

•2022年1月18日

good

創建者 GAUTAM T

•2021年7月31日

good

創建者 Ganesh

•2020年6月4日

Good

創建者 D. R

•2019年11月9日

I'm in week 5, and I think that this course is interesting and you learn from it. However it is done in a somewhat sloppy manner, to my taste.

My biggest problem is the notations and equations are a bit of mess. Beta's in one equation are replaced with phi's in another (sometimes in the same "lecture" slides) or theta's - there's just no real coherent notation. The formulas are brushed through, and they contain mistakes (a product of this sloppy notations), e.g. pi(beta) is missing the beta (which is what it depends on! week5, ARMA properties and a little theory). The R code is also sloppy, for example you see them setting variables in the first cell, and then never using them in the next cell. Or calculating variance using a cumbersome call to an acf function telling it to bring back the autocovariance, and taking the first term. TL;DR - It's just sloppy.

There are no exercises, but the quizzes contain some code you can run. Not enough for really drilling the material into you, though.

In general, I think this course could really improve, and I would like to see it do so. As a general introduction to the topic it might be decent enough.

創建者 Andrea G

•2021年10月31日

This course is difficult as it's not the typical hands on practice with R that you can find anywhere on the web but it goes deep into the math and statistics to let you see and understand what's behind the calls to the automatic routines that at the end we will use. 4 stars are motivated for two reasons: 1. there are some issues here and there with labs and R code 2. I would have preferred a direct-line with teachers and tutors because honestly there are some topics that raise questions and require clarifications but as far as I could see nobody answers questions in the community forum and therefore I found myself going outside of the course to find answers on the web or back to math books which is good if you are persistent but it slows down the overall speed of the learning and the course. So my proposal to coursera : why don't give the possibility (maybe paying for their time) to interact with teachers ?

創建者 Murray S

•2021年4月22日

Judging by some of the comments left in the Discussion Forums, the course name may be a bit of a misnomer. I think the term "practical" conveys more of a hands-on applied focus (using software tools to diagnose and estimate various time series), rather than a more theoretical approach. While there are numerous examples provided, there's also a sizeable theoretical component. While it's certain arguable that setting students loose with software tools and no understanding of the basis of their development is also dangerous, I think "Time Series Analysis" or "Time Series Analysis Fundamentals" might be a better title for the course.

That being said, the course met my objectives. There are a number of links to datasets that are obsolete; it would be good if these were updated, rather than having to spend time tracking them down on the Internet.

創建者 Ron M C

•2019年9月8日

Professors obviously know their stuff and work to outline all the math fairly logical. The title, "Practical Time Series" is a little lost on the actual workload. I am finishing week 3 and I have yet to find anything 'practical' about the course. i'm very intrigued about the math, it is interesting and challenging, but i felt like the discussion in week 1 about all of the data sets we were going to use was a tease.

I would be better able to absorb (not just learn it long enough to ace the quizzes) the material if for each concept there was a practical application of the concept to one or more of the data sets that were made available to us. Because we don't, I often find myself in my own head, searching for applications, and thus not fully paying attention to the videos, which then I have to go back and watch multiple times.

創建者 Matteo B

•2020年5月25日

This is a fantastic course, and I would recommend it to everyone that is interested in Time Series Analyses. After finishing the 5 week program, I can confidentially say that I feel comfortable to start tackling TS projects and build some forecasting models.

However, I deduct one star because the learning curve is very steep and could/should be supported more through graphs and examples, especially in the earlier part. This can be frustrating, especially for people without a very strong statistics background. My best advice for now is to keep going, many concepts become clearer in later lectures.

Overall, this course is highly enjoyable (for a statistics course on R) and I do recommend it to anyone that wants to explore the fascinating world of ARIMA models and time series.

創建者 Jose L A

•2018年11月18日

The course gives really useful skills regarding time series analysis, but it seems a little bit forgotten by the authors since some links in the during the course are not working anymore ( for instance the link describing whether a seasonality is addictive or multiplicative "http://www.forsoc.net/2014/11/11/can-you-identify-additive-and-multiplicative-seasonality/". Also,there are time a future content is presented before the class in some questions, as is the case the moving average week where there is a question regarding auto regressive process, a content present in future classes. Besides those points, the classes and material are really helpful, and i can say that this skills learned will sure be used in my professional life

創建者 Carlos R P G

•2020年1月8日

A very solid introduction to time series analysis, recommended if you have understanding of probability and statistics concepts.

I have seen some complains about the course not being practical enough. The practice comes at the last third of it, and this is as it should be. The SARIMA model is composed of 4 different models (S + AR + I + MA), if you don't understand them independently your chances of doing anything useful with it are slim.

I would have liked it to be more language agnostic since I use Python. The statsmodels module has all the time series analysis tools that you need, and allows to load R datasets, albeit not all of them. The rest you can find fairly easily googling.

創建者 Stefnir K

•2019年12月29日

This course will teach you many of the concepts of time series analysis. It's a good course that is clearly taught by experts in the field and it is no lesser than any course I have taken at a university level.

The problem with him is that the lectures are dry and feel outdated, they are not bad in any way just two professors with a webcam and slides. The second problem is that the tests are a bit easy and you can pass them without understanding by just trial and error.

Overall I recommend this course for those that have little or no background in time series but would really like to dive into this topic. I also recommend at least 1-2 years of Bs in Eng or Sci.

創建者 River B

•2019年11月25日

An excellent introductory course on time series analysis. It has an excellent blend of theory and practice and everything else became intuitive once you studied and gained intuition for the math. Some of the lower rated reviews mention too much theory, but I feel it was imperative to fully understanding the course and am glad they included it. Completing the entire course felt rewarding.

I docked one star because of the sloppiness of some of the slides and equations. Some of the examples don't work either. I enjoy William's videos as his pacing is good, but a lot of times Sadigov tries to rush through the slides as fast as possible.

創建者 Marc-André C

•2018年4月21日

Well built class. I especially enjoyed the inclusion of written material, which I find easier, faster and more enjoyable than videos usually. The material itself is well constructed and the professors are clear. The low point for me comes with the intended audience of the class. At first glance, it is directed toward professionals that have already some familiarity with time series. While I could follow the course independently, I had to rely on other resources to gain intuitions on the concepts. I still don't consider that I could explain the material that I learned as well as I wish I would.

創建者 Derek H

•2020年1月5日

Good introduction to time series analysis, covering the standard curricula of discrete-time stochastic processes, useful statistics, with some additional work with R and some introductory-level theory. The course is not especially rigorous and the quizzes are not hard, but as an introductory course for a person new to time series but with at least partial undergraduate mathematical background, this is a good start. I mostly read the slides to learn the material, as I prefer to read material on my own, and the slides were informative and easy to follow.

創建者 Joseph H

•2021年2月14日

Good introduction to the fundamentals of MA and AR processes, the ACF, PACF, etc. I would have appreciated more applications, more examples showing how these tools are actually used in the real world. Suggestion: Add a week to the course where you take us through several "messy" real-world examples. For example, what are a few key research articles that have used the methods of this course and found REAL INSIGHTS about the studied system? Walk us through a few key real-world success stories of using these methods, from the research literature!

創建者 Jerry H

•2019年1月11日

The course met my expectations, which was to develop basic skills and tools to better understand time series as a jumping off point for some of the work I am doing. I found the practical examples (e.g. coding of the solutions) to be most helpful for my learning style. Also appreciate concept development thrust, to help better understand the applicability and pitfalls of the tools. That being said, I didn't particularly find some of the mathematical derivations helpful, given my bent toward the practical application of the tools and concepts.

創建者 Sian X

•2021年8月9日

The two instructors explain everything crystal clear with cool mathematicians style which I enjoyed a lot!! They definitely deserve 5 stars! The 1 star I took away is because Coursera did not set a good expectation to this class, because of which I'm quite frustrated with how much maths i need to deal with and how less practical it is compared to my 7 years of experience in industry. Though after taking the course i have deeper understanding of the mechanism happening behind R existing functions - thanks to the two great instructors.

創建者 Christian B

•2019年12月29日

I think the course is very helpful and you learn how to perform time series analysis. In the last chapter I was missing the motivation for using tripple exponential smoothing vs. the former SARIMA model. When would I use what? I give only 4 starts due to week 3, where many complained about, and I would agree, that the motivation for the mathematical construction of the Yuri Walker equations is unclear and the lesson itself is a bit confusing. However, week 4 is then way better, when using the matrix notation and concrete examples.

創建者 Anisha S

•2019年5月6日

After coming towards the end of the course, I have changes my perception about the course. It is a great course to learn Time Series Analysis. Though it has some advanced theory and derivations in certain lectures, it has lots of practical exercises as well to perform hands-on. It gives good understanding of time series concepts and different models associated with it. I would recommend this course. I would recommend the instructors to add ARIMAX and multivariate time series analysis in the course as well.

創建者 Joel A

•2020年3月25日

There are some inconsistencies with notation. The quizzes are much too easy and the code/problems are basically given to you. I would say this is a good introductory course to see if you have any interest in the subject. There is some good theory development in the weeks covering autoregressive and moving average models. I did learn some of the terminology and methods for time series analysis that will allow me to go into more developed sources with a basic intuition of some simple processes.