Python for Data Science

Overall a pretty good course and intro to data science using Python. How much you learn from this course is pretty much what you put into it. The grades are very easy to earn and earning a high grade doesn't necessarily mean that you learned a lot. The best thing about the course are the jupyter notebook notes and exercises. They are very detailed and you learn by example. One thing I wish the instructors did was actually grade the projects instead of leaving it to peer review. Peer review is arbitrary and most of the time my peers did not really understand my project mostly because I used statistics which is out of the scope of the course, but I put extra time into figuring stuff out and wish that I could receive better feedback on that aspect. It would have been really valuable if someone could have provided feedback because there was one hurdle I could not move pass despite researching the topic a lot. In my final project one of my peers took marks off just because I didn't have a plot for one of my research questions that specifically asked for parameters that could predict the response well which doesn't need a plot. Meanwhile, I spent my time going through other people's projects and giving feedback whenever applicable. A project in particular I looked at was 400 pages long, 398 pages of it was a long error message. There is also no penalty for giving someone a bad peer review or bonus for giving a good peer review to someone else, so really peer review is a waste of time and most of the time doesn't benefit either side.

Each week of the course takes roughly a few hours to complete. I personally completed the entire course in about 70 hours. 25-30 of those hours working on the mini project and final project, 6 hours on peer review for both projects.

This was my first time taking an online course as if it were a college class. The content is excellent, and the instructors are also excellent. One of them speaks a little fast, but all video lectures come with subtitles. I didn't rate this 5 stars because there was at least one time where I asked a forum question to the course staff, which was also echoed by another classmate regarding potential erratum in the machine learning section that wasn't major, but had me wondering if I had misunderstood some concepts of parallel plots but the edx staff never responded to the question.

Though I spent way more time on this course per week compared to what the course suggests (10 hours/week), I spent most of my time filling up a spiral notebook with notes on concepts, syntax, and other aspects of Python and Data Science to help commit it to memory and to retain as reference for after the course. Additionally I spent much longer on the mid-term and final projects for the course, not because the course didn't prepare me for them but because I chose hard problems to solve so I could get stuck and figure out the solution.

While I have some programming knowledge, I had never used Python before taking this course. I use it every day now in my current job and it has become my go-to tool to solve any data related questions or problems. I would highly recommend this class to anyone interested in learning Python or learning how to think about or process data sets.

I completed the whole course but did not apply for a certificate for the sole reason that I follow MOOC's out of curiosity, not to pursue career interests.
Overall, this course is a very good practical introduction into Python for Data science. While it does not provide you in-depth with the mathematics behind topics such as classification, clustering, etc., it does expose you to the Python Numpy, Pandas and Matplotlib functions so that you are ready-2-go for real-life problems. I wanted to also understand the mathematics behind it, reason why I took the coursera Machine learning course [using Octave] by Andrew Ng in addition. Combined, they form a good introduction into Machine learning.
All-in-all a highly recommendable course with very good teachers.

Overall some (very) good points, and some (very) bad points; that's why at the end I could only give an average mark of 3 for this course.
On the very good side:
- the notebooks: they are of quite good quality. You can learn by example from them. They give you concise efficient learning value by practical example illustration
On the good side:
- the Unix part: I did learn more than a few usefull tips there
- the Numpy part: maybe the best concise overview, and most convincing arguments to use Numpy I ever had
On the bad side:
- the video presentations from the lady Professor are rather weak and not engaging. She is not clear in her explanation, not to say often confused.
On the very bad side:
- no real exercises, assignment or tests except the peer review projects (the test questions are mainly a joke): this is a huge minus for this MOOC. The Professors could/should have spent much more effort building real exercises as application of the course content. They went for the quick and easy way to provide (good) already made notebooks and just explain them. And for exercise, they delegated the workload to external by choosing Peer Review projects for assignments.
I had just completed the R Data Science Specialization from HarvardX before this Course. In contrast, I effectively learnt much more from the HarvardX course due to the much better quality of the exercises and assignments there.

So all in all, this MOOC is ok for the notebook resource. But you likely won't learn much more else. Definitely not worth to pay 350USD to get a Certificate

This tool is the best simplest tool to learn. It has some advantages over other other programming languages.

Python is one of the most popular and fastest-growing programming languages. Inherently, it is interpreted, high-level, general-purpose, and object-oriented scripting language.

Interpreted - an interpreter processes the source file at runtime, it reads the lines of code one by one and performs what is said. Similarly to Perl and PHP, Python does not require that you compile your program before executing it. So, you do not have to invoke a compiler. Instead of running the compiler that helps turn source files into compiled class files, you simply run a .py file. Python byte code compilation is automatic and entirely implicit.

High-level - Python relies on easy-to-read structures that are later translated into a low-level language, the original code that is run on a computer’s central processing unit (CPU). A high-level language is intended to be used by a programmer and the written code is further interpreted into a low-level language. Like C++ or Java, before running, Python has to be processed. This enables Python’s portability — it can run on different kinds of computers with nearly no modifications.

General-purpose - Python can be used for nearly everything. It is applicable to almost every field for a variety of tasks. Be it the execution of such short-term tasks as software testing or long-term product development that involves roadmap planning, Python works well for them all, it is applicable all over the map. Its roles are unlimited. It is popular not only among software engineers, but also among specialists in other fields: mathematics, data analysis, science, accounting, and network engineering. Likewise, Python cliques with young people because it’s a very beginner-friendly scripting language.

Object-oriented - this programming paradigm gives an overall orientation towards scripting and powerful code structuring. This object-oriented approach allows thinking of problems in terms of classes and objects. Then, objects are composed in such a way to make up complex computer programs. Beside the object-oriented programming, Python also supports a procedural paradigm. With OOP being only one of the options, you can make Python programming more advanced by going for an object-oriented programming approach. Developers can create reusable patterns of code thus curtailing redundancy in development projects.

This course is absolutely terrible.

99% of the UNIX coding used in these Jupyter notebooks will not execute on a Windows system. The instructors briefly address that there could be issues for windows users going forward, but don't address how to fix these issues or provide work-around solutions to follow along in the lecture.

The result is I end up wasting tons of time at the start of each lecture, after downloading example files, to devise 'work-arounds' so that I can follow along on my windows system.

I am just auditing the course but the discussion boards are full of 9month+ old msgs from certified learners, who paid to take this course, complaining about these issues and in awe of the fact that the instructors/TAs are nowhere to be found, make no attempt to help windows users, etc.

An interesting, learn-by-doing approach... and course seems to use current packages, versions etc. which is why I give this 2 stars rather than 1. I will continue for a short time longer, in hopes that I might learn something from this course, but I expect I may quit soon.

This was a worthwhile course. I enjoyed it. It was challenging, but not exhausting.

The Python section could have been longer more thorough. I had hoped that the course was more about learning Python, as well as learning about Data Science. Fortunately, this was not my first course in Python.

Peer-review of the mini project was a disappointment. I had taken the task seriously. I reviewed more than the minimum required presentations and I took care to leave comments explaining the scores I had given. The two students who reviewed my work gave less than perfect scores with no notes as to how they reached their scoring conclusion.

Review of the final project was more constructive and I appreciated the reviewers time and effort in explaining their scores.

Hi I audited the course and this is definetely one of the good courses for getting intro to data science with python . Since iam a working professional i didnt had enough time after office completion 6.30 pm , to complete it on daily basis but i sticked to atleast 3 hrs daily .

At multiple places instructor used different ways to do the same task which is good for exposure but confusing for naive learners.

There were few concepts that cud hv been explained more easily.

My advice to the team wud be to explain via a small data set which learners cud visualize and get an idea of and then elaborate concepts on the actual data sets used.

I started with this course around mid october 2017 . I also referred other short you tube videos parallely to reinforce better understanding .thats why i am still in last week of the course in jan 2018 .

But all and all course helped me enabled with the jupyter python broadly and github , conda to some extent .