Good course but with many outdated concepts (mostly valid for Spark 1.x) and some pitfalls. Need many improvements, actualization and some reshaping in the distribution of the topics and sessions of the course. The topics left for the last week are some of the most important and central in current Spark 2.x and they include at least as many fundamental concepts as the rest of the course.

The course content is fine, I did learn about Spark and how to work with it in Scala. The only disappointing thing is the homework assignments all use Spark in local mode, so it runs only on my laptop. I really wanted to see an assignment run in a real cluster, for real. So having some kind of cookbook recipe to follow for putting an assignment into Databricks would caus

I learned a lot and the lectures were good.

The feedback from the automatic grader was sometimes absolutely awful. Cache an RDD in the main -> grader does not execute main -> OOM exception with no stack trace. I guess it prepares you for real life since EMR Spark also has absolutely useless log output if you go OOM.

Thank goodness for forums.

This course took a lot of work to create. I would have like more quizzes during the lectures and the assignments to have walk through the steps more. The best course I have seen online is the Machine learning course by Andrew Ng. https://www.coursera.org/learn/machine-learning/home/welcome.

I did really enjoy the course. Thank you.

Very average. Lectures could be a fraction of their current length, too much time spent rephrasing the same point (sometimes 3+ times!). It was driving me nuts, my mind would wonder if I didn't focus. It would be nice to have local tests that incrementally check progress similar to Andrew Ng's Machine Learning coursera.

The material in this course is very interesting. However, there were a few important issues:

Lots of typos in the slides

Lots of problems with the assignments

At the end, I feel like a beta-tester (it would be OK if it was clearly stated and if we had a discount).

It could also be nice to add 1 or 2 weeks to this course.

I really liked the lectures and the good and fun explanations by the instructor. However, I found the assignments over complicated with unnecessary machine learning concepts involved. I think a course about Spark should be about core Spark ONLY and applications to machine learning should happen in a separate course.

The material of the fourth week is quite dense, this could be split over two weeks (including splitting it into two exercises). The exercise of the fourth week is quite a dissatisfying experience, there is too little detail in the error messages about what failed for students to improve their solution.

It felt like the course material skipped over a great deal of syntax and how-to. It was useful for concepts; but I found that I had to dig a great deal to be able to complete the assignments and that there is a lot of folklore in stackOverflow that potentially send you in a wrong direction.

Explanations are OK and it's a good investment. However, I'm mixed about the courses: the teacher is speaking really fast with slides full of text. It's sometime hard to keep my concentration (compared to previous courses of the specialization). Still, I would recommend this course.

a very basic and shallow introduction:

-spark dataframe usage scenarios and syntax poorly covered

-homeworks are too basic and do not prepare for real life scenarios

-lots of typos in the slides

-pdf slides are not searchable for terms

-how is this course rated 4.7 (as of 24-07-2020)?

The assignments are not really well prepared - there is tests provided which is really needed for big amount of data - sometimes it really hard to find a bug. If you tell me that this is my problem to right tests - I'll tell you that I've paid money for that.

Lectures are great. Explanation are very clear. Assignment was having issue like incorrect and/or vague reporting which made them needlessly painfull. I'm quite sure that the next iteration of this course can get a 5 star rating, but for now, it's only 3.

The lectures are really great with vivid and easy-to-follow explanations on complex topics.

However, the exercises don't seem to match the lectures very well and may confuse you. I would prefer to apply the things I learned in the lectures.

In the assignments, I had the feeling that the goal was more "guess what they want" than "write it correctly using what you have learnt". Stating more clearly the requirements and improving error messages from the grader may help.

Instructor was good and knew what she was talking about. The assignments were also good, but the grading was weird. Spent a lot of time try to figure out the unwritten requirements that would make Coursera's tests pass.

It was a good course, much more useful than the first one in the series. I would say too much focus was on the SQL side and not enough on the big data side, which is what i was hope to get from the course.

I guess it's a goo introduction course into Apache Spark. Meantime I would expect deeper dive into optimizations and algorithms.

There is not any course support at all.

Provided drafts of program assignments does not satisfy SOLID and Clean-code principles

Pretty good, but one of the assignments was poorly set up. Some of the provided code was broken and it was very hard to debug.

Too entry-level after the first two much more challenging courses by M. Odersky.

it's hard for beginner understanding this course

course needs to be updated

a let down and not up to par with other courses in the series.

Huge amount of time is wasted bsically repeating that the API is close to scala collections'. The huge amout of time is wasted again on very simple dataframe APIs including several slides presenting the show() function. Allow me to repeat : several slides presenting the show() function.

Finally the assignments are unimaginative and mechanical. The desciption is really confuse and most of the time is spend trying to chase small differences away to please the grader.

The sole part of the course that seemed interesting was the shuffling one, that was unfourtunately ignored on the assignments

Not related to the course but to spark: What utter mess are dataframes and datasets filled with boilerplate type conversions and runtime erros. I shy away in disgust form this untyped IDE-unfriendly monstruosity.

Relatively decent video lectures, if not that blurry which makes text hard to read. Accompanied with awful practice lessons: - code templates are written with little to no style, even file reading is done in 3 different ways in all 3 lessons; - grader output is very confusing and almost useless; - unit tests, very useful to avoid some common caveats, were present in the first lesson, disappear completely in the last one.

Probably following spark's programming guide is better time investment, even if it misses some "humanity" of video lectures