學生對 IBM 提供的 Scalable Machine Learning on Big Data using Apache Spark 的評價和反饋

TLDR for those who don't want to read through all of that, the course gives a shallow entry into the data engineering part of machine learning. I wished they would make the course more challenging, so that we would learn more.

For people considering the IBM AI engineering specialization and this course, I would say that it is a very good introduction. For those looking for a more in-depth approach to ML and DL, then this course isn't going to hit those areas. Regarding this course specifically, they did a good job explaining the concepts well. I would have preferred if they made the course proejct a lot less hand holding. They essentially give you the jupyter notebook with all the ETL procedures done, and you change like 4 variables, which isn't really intellectually stimulating or challenging. I understand that the course is meant to be an introduction, but I think asking us to do the ETL by ourselves with less rail guards would teach the students a lot more. Like I would say I learned more about Apache Spark and functional programming from the 2nd module quiz than the course project, because the quiz had us writing the code ourselves, and I had to learn and debug functions on my own.

I can't fault Romeo for his enthusiasm and engagement in the forum, and nor do I think his accent is a problem. I can say I learned something from this course, but there are a few negatives

-- Some parts appear unprofessional. This includes the initial videos filmed in the car, the prompts stating that parts are out of date, and the on-the-fly coding in the week 3 and 4 videos

-- The course is initially jargon heavy, but it is pitched at quite a low level otherwise. There is a lot of hand-holding, for the final project you make two alterations to the code already supplied and then copy and paste the results. It would benefit from a review of the didactics.

-- I would have loved to have had more opportunity to play with the data. Why not a tutorial on using SQL or data cleaning? Why not more on the application of the ML tools? There's a definite feeling of being in a sandpit and not being allowed out.

That said, I now have experience with ApacheSpark and I understand how to use it to implement some ML methods, which is good.

The content that is taught was actually satisfying, however, it is obvious most parts of the videos were outdated either due to the fact that they are for another course or they were simply not organized from the beginning. In addition, it would have been awesome If the instructor explained the codes more. I feel that I have learnt the basic idea but I need further self-study to make sense of everything we have covered in terms of the coding.

I enjoyed this class. I worked with Spark a few years ago, but wasn't aware of Pipelines and Parquet. The incorporation of these concepts into the course was useful. The instructor is engaging, but speaks quickly sometimes and there are some translation challenges with his accent. I found myself reading some of the material because i had trouble understanding what he was saying.

I would like to thank Romeo for teaching me. I apologize to rate the course at 3/5. I did like the course in general but I missed the practice of it. The methodology process did not help me to learn the practice. I scored better in most quizes on the first attempting while I could not guess how the code are written. I wish I did learn to interpret or rewriting the code

Regards

Its a nice course and good way to start Apache Spark.But I feel its a bit too fast as well as too high level for those who are pure machine learner and deep learner practitioners on jupyters and colabs,they are gonna find it bit tough and programming part will go over the head.So Goodluck.

But its a nice way to start learning a fascinating technology of Apache Spark.

I really enjoyed this course, I think despite this it requires a review to make it more consistent. One thing that would have made it better for me. Would have been to have the jupyter notebook matching exactly the video so I could have worked on them in parallel... Thanks very well build course, it really gives a good base to start using Apache-Spark ML

The material presented in the course is important for everyone looking to go into the Data Science or Machine Learning fields, but some of the examples in the earlier chapters use Python 2 and have not been updated to Python 3. The learner has to go hunting themselves in the forums for official posts on how to fix these error (they were there).

It is a very interesting course. Some videos and lectures however should be improved:

-start with a purpose: what is the goal of this script? What do we want to learn from the dataset?

-the explanation of the sliding windows was a little bit obscure.

The scripts are useful and if the learner plays around with them she/he can learn a lot.

Although this course introduced us to the very important idea: distributed and parallel processing, but I find it too broad and too high level. We didnt go deep into any of the topics, and the assignments are to easy(some of them are already done, you just have to find the correct number for the outputs and place it in quiz section)

Interesting, but not much opportunity to practice what is taught. Instructor walks through a lot of examples, but they are hard to follow because his notebook screen is a bit blurry. A lot of type a long, and trust me, or "we will get to this latter". Pretty easy compared to other similar coursera courses I've taken.

Found this course difficult compared to others, as i am a mechanical guy. However, resources provided in this course are great. In this course unlike others requires lot of reading from resources. Finally, enjoyed this course. Only thing that troubled me is the instructors slang of English -) which is my problem not his.

Like the format of this course, which seems more laid back. Having said that, some of the assignments had some confusing portion, but need to acknowledge this is an intermediate course and not a beginner one. I also missed the part of the explanation that Apache Spark has its own tools vs. using Python's SciKit

Sources in the lectures were really great to understand what is Apache Spark and How to use it.

However, in some part of the lecture, I loss my way to understand what's going on here...

Anyway, at final course, I could review what I learned in this course and that will be a good guide to use Apache Spark.

Slides contain some typo in Python codes but highlighters are available to let you know what are wrong. However it still makes me feel that the course materials are not very well prepared.

Good thing is thing I have got a basic understanding about how Apache Spark can facilitate machine learning.

It will be nice to have a better video quality, since we cannot really see what's the instructor is typing in the video. Especially that the syntax is changed (from Python 2.x to Python 3.x) and Apache Spark also have new version (3.1.1), many of the contents can be updated using latest APIs.

Very interesting course, learning about utilizing Apache Spark parallel processing and how to build ML models. Video quality was not satisfactory for viewing the described Python code and I had difficulties understanding the spoken language, fortunately the video's transcription helped.

Great course. There should probably be more coding tests where submissions get you a grade like some of the other Coursera coding courses.

Some of the coding in the lectures is a bit too quick, but that's probably just for because I have never used the Apache Spark syntax before.

Very good teaching techniques, The professor explained everything well, The sound quality was dull on 2nd week's video and the accent was a bit tricky for me but the quizzes were good and if you code with him you'll be able to understand the concepts easily

I like the course, but it fails to mention clearly how learning apache spark could help us. Also, it requires a certain amount of coding experience, I was able to finish it, but sometimes I had no idea what I was doing.

The course was both concise and specific in the information provided, covering all thhe key points of the various concepts. Only aspect I found it slightly lacking in was the organization of excercise material.

It was surprisingly fast-paced. There were a few intuitive leaps, including a bad data reference on the final project, that were potential stumbling blocks, but I feel more confident having overcome them.

It is a good course for beginners in the domain of Apache Spark and Apache Spark ML. Programming assignments could have been better if they were applied to "Big Data" and not on toy datasets.

Again a nice course that introduce you on Apache Spark Usage. Just a little suggestion, if you could insert a little tweak on how pass from spark to pandas and vice versa.

Enjoy :)

Teaching was clear and understandable. Only feedback would be I hope the lab work would be more hands on because I'm worried I don't pick up the concepts unless I type them out.