Chevron Left
返回到 大规模数据处理:系统与算法

學生對 华盛顿大学 提供的 大规模数据处理:系统与算法 的評價和反饋

748 個評分
162 條評論


Data analysis has replaced data acquisition as the bottleneck to evidence-based decision making --- we are drowning in it. Extracting knowledge from large, heterogeneous, and noisy datasets requires not only powerful computing resources, but the programming abstractions to use them effectively. The abstractions that emerged in the last decade blend ideas from parallel databases, distributed systems, and programming languages to create a new class of scalable data analytics platforms that form the foundation for data science at realistic scales. In this course, you will learn the landscape of relevant systems, the principles on which they rely, their tradeoffs, and how to evaluate their utility against your requirements. You will learn how practical systems were derived from the frontier of research in computer science and what systems are coming on the horizon. Cloud computing, SQL and NoSQL databases, MapReduce and the ecosystem it spawned, Spark and its contemporaries, and specialized systems for graphs and arrays will be covered. You will also learn the history and context of data science, the skills, challenges, and methodologies the term implies, and how to structure a data science project. At the end of this course, you will be able to: Learning Goals: 1. Describe common patterns, challenges, and approaches associated with data science projects, and what makes them different from projects in related fields. 2. Identify and use the programming models associated with scalable data manipulation, including relational algebra, mapreduce, and other data flow models. 3. Use database technology adapted for large-scale analytics, including the concepts driving parallel databases, parallel query processing, and in-database analytics 4. Evaluate key-value stores and NoSQL systems, describe their tradeoffs with comparable systems, the details of important examples in the space, and future trends. 5. “Think” in MapReduce to effectively write algorithms for systems including Hadoop and Spark. You will understand their limitations, design details, their relationship to databases, and their associated ecosystem of algorithms, extensions, and languages. write programs in Spark 6. Describe the landscape of specialized Big Data systems for graphs, arrays, and streams...



Great course that strikes a balance between teaching general principles and concepts, and providing hands-on technical skills and practice.\n\nThe lessons are well designed and clearly conveyed.


I like the breadth of coverage of this class. Each of the exercise is a gem in that I get to learn something new also. I would highly recommend this even to experience practitioner also.


126 - 大规模数据处理:系统与算法 的 150 個評論(共 158 個)

創建者 Martin M


Good content for Data Scientists but video lessons are not sufficient to be able to complete the assignments. It required great deal of own searching and trials and errors to complete the course.

創建者 Ingo B


This is a cooked up version from an earlier, more extensive course. Lecture videos now split from 10-14 minutes into lots of 4-6 minute videos. It seems, some assignments are missing, too.

創建者 Dwayne B


Good information but lectures were poorly produced and unedited and exercise instructions were blatantly incorrect several times.

創建者 Andrea R


A couple of comments in the forums were very old and it seemed nobody had been checking the course for a long time.

創建者 Ryan S


Long, slow, rambling video. I watched most of it at 1.75x. Slides are kind of a mess and lectures are disorganized.

創建者 Fisher


little touch of everything, it's good intro for non-tech, but way too shallow for a student from tech background

創建者 Griffin S


Program instructions could be more specific. Make it clear exactly what format the programs output should be.

創建者 James S


The material is good. If you can get past the instructor's mumbling and rapid speaking then you'll be okay.

創建者 Tushar T


Assignments were just not that challenging except first one

創建者 Daniel V


If you don't know Pythonl, don't take this course.

創建者 Maria K


Interesting but outdated

創建者 Ian P


This course, which sounds promising in title and syllabus, has many glaring deficiencies. In fact, I feel terrible if anyone ponied up $100+ for it. It roughly covers some concepts of data science, but never at scale, and never very clearly. My background is a science Ph.D. with a lot of computational science experience.

The lectures: Clearly poorly planned. Bill Howe has some knowledge about databases, but little skill in communicating it. The organizational structure leaves much to be desired. Much of the lectures are broad-brush and halting, simultaneously being too detailed as times and not broad enough at other times. Technical portions are marked by a number of errors in speaking and on the slides, as well as a lot of hesitation and jargon. It's as if he neither thought about the structure of what he wanted to say or a script of what he might say prior to recoding the session. Phoning in it is an apt description.

The Assignments: The first assignment with Twitter was fun and interesting and gets the course 2 stars instead of one. The lectures prior to this will not prepare you for the assignment though, so might as well just skip them and do it on your own. The SQL assignment followed a set of lectures in which no proper discussion of SQL was ever given. The last assignment on Map-Reduce is acceptable although a number of errors in the homeworks are still uncorrected long after the first offering of this course. The autograder's idea of helpful feedback is similar to "Incorrect value. Try again" Week 4 of this course, which contains a vast amount of information has no exercises at all.

Overall, this class is the polar opposite of a quality online course like Andrew Ng's Machine Learning Course. Do the twitter assignment and skip the rest. Lectures are poor and assignments are well below average. If I were at UW, this is not the kind of course I'd want representing my university in a public setting.

創建者 Alon M


This course offers a mixed learning experience:


(1) some lessons offer deep understanding of SQL and MAP REDUCE algorithms.

(2) the HW is challenging and gives you an opportunity to actually implement SQL, DATA MINING, and MAP REDUCE. (con: those areas are big, and this course only slightly touches each)


(1) the tasks are very hard, simply because they are written bad(!!!!) : there are discrepancies between the written task and what the Automatic grader checks, insufficient documentation, weird way to submit points among the problems, weak help from mentors (if at all). boy, I found myself pulling my hair out.

(2) this course takes it for granted the you already know SQL, and how to work with linux based OS in a virtual box, and of-course: python is a must. If you are new to one or more of those subjects, this might not be the course for you.

(3) the talks offer very little help in solving the tasks. so basically you are on your own and need to search for clues in the web or read out comments from other desperate learners.

創建者 Marcio G


This course is quite outdated. I didn't learn much beyond what I already knew before I started. The Spark courses from edX are way better than these. Hopefully "Big Data Analysis with Scala and Spark" from the "École Polytechnique Fédérale de Lausanne" (also from Coursera) is good (I know their Scala courses, which are taught by Martin Odersky, are quite good).

There are very few quizzes between lectures and the assignments are not very challenging.

Many of the videos, specially the ones at the end were extremely rushed over. They serve more as a review if you know the subject, otherwise I don't think most people will get much from them.

The audio isn't very good for most of the lectures, many having an very annoying chirping sound (from when you leave an old flip phone near a computer... "teh-teh-teh teh-teh-teh teh-teh-teh teh-tehhhhhh....". Gosh, I haven't heard this sound in maybe over five years...).

創建者 Coen J


Good focus on ideas vs principles. The focus on relational algebra is a great way to look at data manipulation in general. Unfortunately, relational algebra is explained quite well, but not really applied after that. This could be a great course if it really taught to constantly think in terms of relational algebra.

Okay-ish explanations of databases and hadoop. Not very deep and not always structured, but rather focused on the technology principles instead of the data principles.

I think that this specialisation suffers the same problem most data science/mining/analytics courses suffer: it ignores the non-technical starting point: scientific or business relevance. How does one organise data, get to know completely new data, understand possible value? i.e. how to start a data science project if all there is is unorganised data and the wish to do 'something' with it.

創建者 Ganeshwara H H


1. The title is misleading since "at scale" led people to think that large scale data processing platform such as spark and nosql databases will be central to the course right from the start

2. The assignments need a lot of improvements. I am not happy with how we're often only required to submit a single number as an answer. The biggest problem is that this way the grader won't be able to give you meaningful feedback / hint of where you might be wrong. A grader that only tells you "your answer is incorrect" does little to help you learn from mistakes.

3. I think assignment 2 can be paced differently - now it feels that we have a bunch of very easy parts (a-g) that is not very interesting, where the last three are significantly harder.

創建者 Andre J


I'll say the same about this class as the rest of the specialization, if you have the skills to complete this course then you don't need to take this course. If you don't have the skills to complete this course, you will not complete this course. The course instruction is at 10000 feet level and the assignments are very challenging and the course will NOT teach you the skills required to complete the assignments.

I recommend the Machine Learning Course (from Bill's colleagues) at University of Washington. That is a course where you get some real instruction and understanding of how to complete assignments (though still very challenging).

創建者 Igor S


This course left me with mixing feelings. I learned some new things, but overall I don't think that I got understanding of base concepts. Week 4 seems to have disproportionately more material than previous weeks, as though authors tried to Although free, this is course is also offered as a part of paid specialization, and I would be really disappointed if I'd spent money on a course like this.

創建者 Hongmin C


The grader is horrible leaving you with such brief error messages. You never know what is wrong with your code. The forum is not as useful as I expected nor as it is in other Coursera courses. You know, few classmates. The videos provide you with tons of information, but not much of them are well-organized. I often felt tired and confused since these long videos seldom got to the point.

創建者 Ben K


This course probably deserves 3-4 stars in a better, maintained form, but the entire specialization is not maintained, the lectures have no production values. Basically, it's a money pit that Coursera is keeping up cynically. It's a real shame because the syllabus correctly addresses a gap in most data scientists' skills.

創建者 Supharerk T


The exercises are fun and challenging. However, the lecture are not related to the exercises and are very hard to follow (I think it's the same thing as Brian's class in Johns Hopkins' data science course) If you are taking Bill Howe's class, just go straight to those exercises and skip lectures.

創建者 Diego P


Many mistakes in the slides and poorly defined problems in the assignments have gone uncorrected for over a year. The content is very basic, as would be for an introductory course, but can even serve as a refresher for CS graduates.

創建者 Jana E


Quite interesting subjects, but video material is not of high quality and many mistakes are not changed in later sessions but altered via a text in the screen of a note on the next sheet.

創建者 Lei Z


The course is good. But it does not has lecture slides that is better for students to understand.

創建者 梁司其


boring and easy, the homework is too easy and not well designed