뉴욕주립대학교의 Practical Time Series Analysis 학습자 리뷰 및 피드백

Contents of the course are very good. It provides a compact introduction to the most important statistical methods for time series estimation. Nevertheless, there are many mistakes in slides/exercises, necessity to download datasets instead of using predefined ones, and a great number of time series libraries to download when we could just use 2 of those, which is ok, but I took 1 star for it. The other star taken was because I think the course could still include more information on the theory of time series (yes, I understand that the name is PRACTICAL time series, but still), and more meaningful exercises. Most of those were pretty trivial.

Overall, it is a pretty good course: in just a few weeks you are introduced to a satisfactory level of time series analysis with R even if your background in R and/or statistics is not that high. I have personally missed some more depth (even if optional: for example, providing additional optional readings, or creating an honours track with extra difficult assignments for motivated students). Instructors are pretty good but they should make an extra effort to be clearer and not to speed up too much: some more additional videos or just more time for each video covering the same content would be better.

While this course was pretty thorough, it was way too mathematical and not nearly as applied as I would have expected a 'practical' course to be. It was also extremely frustrating that a large proportion of the forum messages went unanswered - I've never done a Coursera course before where there hasn't at least been a research assistance answering questions. You really had to try to follow everything yourself, and if you had a question, you had to find the answer yourself. A bit disappointing.

It was a nice course, very informative. But week 3 simply had way too much matter in it in comparison with the previous 2 weeks which makes it difficult if you've set a timetable. Also, week 6 had poorly maintained data availability, and missing code. Lastly, the professors are absent from the discussions. There are unanswered doubts from longer than a year ago.

But if you dedicate enough time and effort, this course is pretty good.

It's decent for what it is - a course trying to balance theory with practicality. However, it uses R for the target implementation environment. That's fine, but as a matter of taste, I prefer python. Therefore, for me, the practicality aspect of the course wasn't the greatest, and that was its main selling point.

It is in R, making it rather limited for applications in industry, however using Python's statsmodel you should be able to keep up with the notes.

I was expecting a lot more theoretical explanation of the concepts rather than just demo'ing some R code and fitting some libraries to data.

It was definitely a gentle and accessible introduction to time series analysis. I was hoping for something a little more comprehensive and challenging, however, and sound quality was sometimes poor. This would serve as good preparation for an introductory course on time series.

Overall, this is a good course for beginners in time series analysis.

However, I wish they had go beyond the basics. I missed more advanced content, discussions between different forecasting techniques, multivariate forecasting...It is pretty basic.

Very interesting course. For a 'practical' course, I did find it rather theoretical. Also, some links to downloadable data are missing and there are some small errors in the R scripts that are offered. Nonetheless a good course, I learned a lot!

The coding for the steps confusing and can follow how the tutor explained. I would like to suggest for the course coordinator to create a document on steps with coding for each model or test we want to run. Thank you.

Good/thorough introduction to moving average, auto-regressive, intetgrade, and seasonal models as well as basic forecasting. Some of the labs don't work, as well as links, and the organization is messy in places.

While professor Thistetlon gave the classes in a way that kept my attention,Prof Sadigov wasn't so good at it. Clearly he know a lot but not as a teacher.

Would have loved more work with the ready made packages, I feel like I still can't master using them in R, although I did understand the material.

good, but not great. needs more explanation about the intuition. needs a lot more exercises. needs to cover more topics.

Instructors are good but as per the title this is not more than 30% practical Time Series Analysis

The content was great, but I'd have liked to see more practical exercises.

I find the course a little too theoretical

I finished this course with grade over 90% and yet I feel disappointed with it.

Divided between intense theoretical (I could say "impractical") formulations and forcing implementations in R, it feels like the course could never really achieve the objectives of teaching what it was meant to.

And so, I feel like I will hardly be able to implement any of the given topics by myself with my current knowledge...

Apart from my critique of the course, the course really needs a revision because many contents are not up-to-date (like filenames) or even working (several pieces of code throughout the course, specially crucial in quizzes)...

Not as practical as I expected

The title and prereq's to don't reflect reality.