Very good course

Good!

k

I found this course to be very approximate as if it just wanted to give a high-level idea of the concepts it covered. But maybe, that is the actual goal of the course: give an idea of how ML concepts can be applied to the finance domain and then let the student deepen and practice with the techniques shown. The parts that I've found to give more interesting, even though they have not been covered in detail, are the quant strategies and the time series one. The ML part, coming from an ML background, is well explained but they have been formulated only to give a high-level idea without going into the mathematical details(which I think it's outside of the scope of this course). Regarding the lab part, I didn't enjoy the BiqQuery part while I've loved the lab with Jupiter notebooks (I'm a little biased here). I would have liked more math details, but again that is just a personal preference.

First 3 weeks were quite good, although I found lack of lab practice. The time limitations on using GCP account were slightly pushing to complete it fast without having time for thorough thinking and experimenting. Although they could be restarted - the work had to be recreated again when this happened. Last week was very shallow and non-consequent and looked like it should be the first week as there were explanations of ML and GCP AI Notebooks. Which had been used during already during the first 3 weeks. Although I'm impressed with GCP platform and its AI capabilities, I felt like it had been highly advertised and selling though the course, where my personal preference would be learning more of algorithms and experimenting and using GCP just as one a tool.

Most of the course is used to advertise GoogleCloud, in particular BigQuery, instead of teaching more general concepts. At least now I know what BigQuery is and, as a python programmer, I won't be using it since I don't want to learn a new language used and managed only by Google. The AI notebooks are awesome though! Exactly the opposite of BigQuery: uses a language and tools that everyone knows (python notebook) but on a virtual machine managed in the cloud. About the teachings, I learn a bit more about time series, but just the tip of the iceberg really, nothing applicable. Looking forward to the real deal, opefully in the next course.

The course isn't really for complete beginner. It requires additional readings and googling on your own to understand the gaps. The labs are great to provide hands-on application (albeit requiring some knowledge of Pandas, scikit-learn and statsmodels) and but I feel that some of the content could have been discussed more in details in the video lectures or as supplementary readings. Nevertheless if you are willing to spend extra time researching to understand better, this course is still great for you.

I found the course introductory, not intermediate. There is no logical connection between the GCP part and its direct application to trading. Simple Notebooks would have been a better choice. The examples are pretty rudimentary but you can get some ideas here and there. The last week covering Deep Learning does not touch anything at all about Trading.

Not a memorable course, but at least it is lightweight. The Google part does not show any value in my opinion, it is like two independent courses in one.

Pulling the lessons and homeworks from other courses was probably necessary to keep the costs down and to make sure everyone is on the same page with the tools and procedures, so you're forgiven for doing that. I am looking forward to the next class so we can hopefully build on ARIMA and go beyond it to explore correlations between multiple time series and also how to do feature engineering that takes into account multiple time scales simultaneously (min, hour, day, week, year).

Like the finance aspect of things but the machine learning part seems to be pieced together from other lecture series and doesn't have great relevance to the trading topic. Also I am not here to learn the way Google do things albeit its power. You can run a separate ad channel if you want but why Coursera?

The last assignment cannot be done in Hong Kong as the Jupyter cannot be opened using Singapore or Australia server.

Also, the AI platform has been upgraded by Google Cloud and it is now Vertex AI and not the AI platform before. The look and the button are placed in different areas.

Very well covered the theory of Training and its representation as a problem solving tool when it comes to predicting future prices, however the Machine Learning Part is not at all integrated to the Specialization (Trading) and also its content is not interconnected!

This course tries to be an introduction but it's not enough to give you an overview of ML and not well prepared also.

Labs are not encouraging to solve problems, it gives very trivial problem without focusing on enhancing your abilities on the learned subject.

Useful as an introduction but feels like a patchwork of components that have not been developed in a consistent manner. For example the week 4 material is by and large a replication of a subset of what is found in the "How Google does AI/ML.

the platform they chose for submitting homework was not very well. most of the time it took me 15min to get to the platform to code the homework. also, it doesn't dig deep into topics, which is fine because it is a comprehensive course.

The course has a few good points, but the lectures are often superficial. Also it seems like they are taking lectures from different courses and assembling them into one class, in a way that the lectures often seem disconnected.

Decent intro for ML but very limited in how it relates to trading. I would not say I feel comfortable creating an algo after this course. Also, felt very much like a Google ad quite often.

A decent introduction but with minimal hands on learning. Most of the initiative is on the student to learn about and apply the topics described here. Also tends to promote GCP a lot.

The course seems incomplete or not organized well. The concepts come and go out of nowhere. I knew lot of concepts beforehand and could figure out, but it was very fragmented.

The course was mostly a series of lectures and not recommended for absolute beginners in ML and BIG Query ML. According to me , the course was mildly satisfactory.

The course is loosely organized. Some of the concepts in lab have not been gone through in the lecture. Anyway, the lectures presented by Jack is good.

Theory of the financial part was interesting. "Coding" part was more or less running some fully implemented scripts from a github repository.

The course seems to be more focused on advertising google cloud platform and there is hardly any focus on how to use ML or AI for trading.

Good introduction but would be better served with labs that force you to code more rather than do execute pre-existing commands

Too layman. NO real ML techniques for real trading, just some intro that you can easily find on stackoverflows or something.