学生对 密歇根大学 提供的 Applied Plotting, Charting & Data Representation in Python 的评价和反馈

good

good

Cool

good

Good

Good

good

good

good

Nice

good

good

Good

good

Good

nice

nice

pop

L

G

/

This was a very good course and I feel like I really learned a lot. It starts out with a module on how data visualizations can mislead (intentionally or unintentionally!) and principles of good data visualization. I found it very informative. The instructor used good examples and also walks through the tweaking of a plot to follow the principles of good data visualization using matplotlib.

Subsequent modules give you lots of experience with the matplotlib library and the process of transforming data and making visualizations. There is often a moderate amount of flexibility in the assignments, giving the learner the opportunity to take as much as they want from the experience. You get back what you put in. Multiple assignments have you go out in the real world to look for charts and data to use in your assignments, which give you a very good opportunity to apply what you've learned, moreso than in constrained assignments where data is given to you.

My only complaint is the peer grading system, but I wish I could only dock half a star for it. I think it was actually implemented very well for a peer-grading system in a MOOC, and I even think that peer grading was the appropriate choice of grading system for this course, but of course it is going to have its downfalls. You are graded on a rubric provided by the instructor which mostly awards points just for completion of various parts of the assignment. This is how it has to be, as judging the quality of a visualization is very subjective (and dependent on the grader's comprehension of the course content), and students' grades shouldn't suffer because one of their graders doesn't like the colour red or because they don't understand the principles we are supposed to be applying.

As an aside, there are many complaints that students are directed to read the documentation of a given library, google, search stackoverflow, etc. A course cannot teach you everything you need to know about a library. You are going to have to look things up yourself, especially when debugging. This is the reality coding.

I liked this course. It was very practical. The reading material was very interesting and valuable. Project had some flaws in terms of formulation of questions but they really pushed me to search and look for help. That in my opinion was actually good and I think that after each of these courses there is still a lot we need to learn ourselves. The course materials was laid out pretty well. Sometimes I had a feeling that things were a bit rushed (like introducing seaborn at the very end) . There is a lot of software terminology involved in the beginning when explaining matplotlib notebook and it was hard to follow that part. My biggest complaint is the Coursera's online Jupyter notebook. I got very frustrated with it. Even if you save your work frequently it often happens that the connection does not work or isn't strong and after you close your assignment you edits are lost!!! My advice, download all necessary links and work in your own Jupiter notebook and then upload the assignment.

I think the greatest strength of this course is that at the end of this you will be very confident in writing code for creating data visualizations. However the expected timelines for completion of assignments are completely above expectations. Unless you already have experience with matplotlib and its API, it is difficult for students to cope up.

But on the other hand searching for stuff online on stackoverflow and matplotlib also was a really enlightening experience and teaches you are not the only one having these kind of problems. I would have loved to see more video explanations on ScalarMappable objects which was a huge part of the assignment but was not covered in the video lectures. Also would have loved to see more concepts explored about sea born package.

Great course!A couple things keep it from being 5 stars. 1 - the content comes a little fast without enough reinforcement. The balance here isn't perfectly struck as it is in the 3rd course of the specialization - Machine Learning. Although the content of week 1 is good, I think quite a bit of it should be optional and substituted with more coding exercises and longer assignments in the subsequent weeks. Week 2 has a bit too much esoterica for an intro course, and I'd rather have week 3 and 4 concepts reinforced more instead. At the end of this course, and after a few days, I'm confident I can look back and make any kind of plot I want. A minor quip - much of the code for the course will throw deprecation warnings in the latest versions of matplotlib.