Data Science Courses/Degress/Work

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi,

I was the Director of a famous AI / Data Science firm. I spent some time working as a Data scientist in the US prior to this. I also took Masters in Data Science in an ivy league. Currently, I run my own firm that deals with artificial intelligence.

Didn't see a thread that is dedicated to data science so I thought it would be a great idea to create a thread dedicated to this purpose. This is just my little way of giving back, if anyone needs any advice in areas such as:

- Data science courses in Singapore
- What overseas masters are worth it?
- Working experience in US (data science)

Please feel free to chime in and let this thread flourish. Cheers:)
 
Last edited:

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi,

It really depends on your goal. If your goal is just to gain some insights on how data science can be applied to solve business problems, I think it would be able to address that to some extent.

However, if you wish to be a proper data scientist / data analyst, then it definitely is not enough. Generally, the poly data science courses lack the required rigour, and the lecturers are also not really experts in this field. The lecturers are usually some mechanical engineers with some programming knowledge. I speak from experience here, having interviewed several poly lecturers for a data science instructor role at the Academy which I was heading.

Anyway if you are serious about a career transition to this field, consider:
1) General Assembly (data science immersive)
2) Heicoders Academy (it's a budding academy, but I seen their content and their instructor profile, and it is solid).
3) You can also consider some of the programs offered by (NUS & NTU). I would advise you to steer clear of SMU's analytics program though, they aren't really strong in this field yet even though they were the pioneer.
 
Last edited:

Pianoviolinlover

Senior Member
Joined
Dec 9, 2018
Messages
1,722
Reaction score
1
Hi,

It really depends on your goal. If your goal is just to gain some insights on how data science can be applied to solve business problems, I think it would be able to address that to some extent.

However, if you wish to be a proper data scientist / data analyst, then it definitely is not enough. Generally, the poly data science courses lack the required rigour, and the lecturers are also not really experts in this field. The lecturers are usually some mechanical engineers with some programming knowledge. I speak from experience here, having interviewed several poly lecturers for a data science instructor role at the Academy which I was heading.

Anyway if you are serious about a career transition to this field, consider:
1) General Assembly (data science immersive)
2) Heicoders Academy (it's a budding academy, but I seen their content and their instructor profile, and it is solid).
3) You can also consider some of the programs offered by (NUS & NTU). I would advise you to steer clear of SMU's analytics program though, they aren't really strong in this field yet even though they were the pioneer.

Even SMU prof cert in python programming?
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi,

You are talking about the $10k++ course which teaches the following right:

LEARNING OBJECTIVES
Upon completion of the course, participants are expected to have acquired a working knowledge and appreciation of:

-Fundamental concepts in programming
-Use of functions and data structures in Python
-Object-oriented programming paradigm
-Basic data wrangling in Python
-Applied data analysis using Python

(Copied from SMU website)

If so here are my comments:
- The curriculum does what it promises, which is to teach you python. But nothing more. It doesn't delve very much in data analytics. So if you are expecting to be able to work in a data analytics role after taking this course, that cannot be further from the truth

- This course is very expensive for very little content covered. Heicoders Academy and GA covers most of that content in one course. For Heicoders, with a $2.4k course (see www[dot]heicodersacademy[dot]com) they already cover all that content (minus OOP). GA's course does the same, but I think it is more expensive. I don't have access to GA's latest cost unfortunately. You will have to email them yourself since they don't have the cost on their website (generalassemb[dot]ly/)

- The good thing about the SMU Academy course is that the qualifications of the instructors are good. I went to the same overseas school and same course as some of them, so I know they definitely have what it takes to teach. But unfortunately, I feel that the curriculum is very lacking. Perhaps that is due to the fact that the instructors are more used to teaching kids (Most of them come from SGCodeCampus, a programming school that focuses on kids). So data science isn't really their expertise.

So overall, I don't think it is bad. It is just that the course is really not value for money. And it doesn't train you to be a data analyst or data scientist.

Today, I just got off a call from a Data Science Tech-lead in a top US company today. And she was sharing with me that the expectations for entry data scientists is that they can at least do the following:

- write supervised machine learning models
- have a strong grasp of statistics (understand probability distributions, and basics like skewness, kurtosis)
- have a good mastery of data wrangling process (done through tools like Pandas)

You need to look for courses with keywords like the above if you seriously want to get a role in Data Science.
 
Last edited:

Alsonlow32

Junior Member
Joined
Aug 11, 2020
Messages
7
Reaction score
2
Hi, what about courses from Mages Institute? I understand that they have career support upon completion of the course. The course fees is around <$6k
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi,

Datacamp's materials are pretty decent and very holistic. But honestly the limitation to online courses is that they seldom cover about industry applications which face-to-face instructors can. For example, they don't teach you about code hygiene, or how to do proper documentations.

This might seem like a small thing, but this kind of practices is highly valued especially in top tier tech companies. So I would say Datacamp is good for building your data science technical skills. But when you want to learn how the industry operates, and how to apply your skills in the industry, having a real-life instructor is more advisable.

That said, I use datacamp to refresh myself or learn new techniques / stuff. For the more advanced material, you can find them at KDnuggets. As for dataquest, I find it a bit too amatuerish and not very well organised. As for Codecademy, I have not taken enough of their data science courses to give a fair comment.
 
Last edited:

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
I really don't recommend Mages if you want to take Data Science. For game development, then yes, they are one of the best. But for data science, please save your money.

I once interviewed and hired an instructor from Mages before for a part-time instructor role in an academy. He claimed to have 30++ years of experience. And you know what? His skills in python was seriously lacking (this was my fault for not imposing a technical assessment on him. I was 'fooled' by his profile. He graduated with a degree in com science in the same university as the guy who popularised Neural Networks. And taught in poly & Mages. I learnt that lesson ever since).

And his class was very boring (received teaching evaluation of 7/10 from students, one of the worse ever).

So yea, I really don't think very highly of Mages Institutes instructor when it concerns courses beyond game development.
 
Last edited:

b0ring

Senior Member
Joined
Aug 9, 2019
Messages
585
Reaction score
4
Thanks for the inputs boring!
Forget to add also, for general assembly data science immersive course, you must know python programming before they will accept you. If not they will suggest you to take their python programming at around 5k plus as well or you need to go learn python somewhere yourself. There will be some pre assessments to do as part of the requirement to be accepted into the course.
 

parchiao

Arch-Supremacy Member
Joined
Aug 1, 2003
Messages
13,333
Reaction score
5,612
TS. Can you list out the maths, applications and languages needed in data science. Maybe you could share which are more commonly used by what sectors. I have an awful terrible stinking understanding of it. But I do have a maths plus programming plus data background which I have never put to good use. :)
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
TS. Can you list out the maths, applications and languages needed in data science. Maybe you could share which are more commonly used by what sectors. I have an awful terrible stinking understanding of it. But I do have a maths plus programming plus data background which I have never put to good use. :)
Hi to list some:

Math
1) Linear Algebra
2) Linear optimisation
3) Probability models
4) Statistics (need understand distributions, hypo-testing, 5-number summary etc)

Languages / Technology
1) Python (MUST be strong! Make sure you are very well acquainted with the pythonic way of coding such as list comprehension)
2) Pandas (Another must! My good friend who clinched a >10k/mth data science job in US. He was tasked to just do data transformation with pandas for a few months. If you suck at Pandas, no one would trust that you can do anything else)
3) Visualisation (Have a visualisation library you are good at. Seaborne, plotly, bokeh)
4) Machine Learning Models (At least know how to implement this via Sklearn, but if you are good, please go and learn how to do tensorflow)

Techniques
1) Dimension Reduction
2) Hyperparamter tuning
3) Model evaluation (ROC, MSE)
4) Feature Engineering (super important)

Bonus
1) Spark / Hadoop (Many companies are still in the midst of learning how to run models on distributed computing systems, so knowing this would be a great bonus. I guarantee you that if you are good at this, you will definitely have a job)

All that said, please do not walk away with the idea that fancy pansy models are always the best. They are NOT. Often times, the ones that really give you business insights and results, are well implemented simple models like your linear regression & logistic regression. Some of the best housing pricing models are linear regression models. So any course you take, MUST also teach you the intuition behind the application. I feel very strongly about this, and so do my hiring manager friends for Data science.
 
Last edited:

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Forget to add also, for general assembly data science immersive course, you must know python programming before they will accept you. If not they will suggest you to take their python programming at around 5k plus as well or you need to go learn python somewhere yourself. There will be some pre assessments to do as part of the requirement to be accepted into the course.
Haha yea you are right Boring.

I just want to add on 2 caveats about General Assembly:
1) Don't expect the same career support you will get in General Assembly (US). The one in US is so much better, because they are very deliberate in sending your resumes to their career partners. I know of many who taken the Data Science immersive in GA (SG), but they were not given any career support at all

2) The good Data Science Immersive Instructor I mentioned - Misrab, has already relocated to UK. So, I am not very sure of the current quality of their instructors now.

As of this moment, the only school I can vouch for in terms of instructor quality is Heicoders Academy, and maybe Metis (I interviewed one of the guy before. He is not the most amazing talent I interviewed, but he has the right idea and approach for data science)
 
Last edited:

Triumvirate

Senior Member
Joined
Oct 8, 2019
Messages
2,152
Reaction score
405
Oh man, I'm enrolling in either SP or NP's data science socialist diploma, I thought there were useful certifications to have. Any advice on where to get good and recognizable certifications, online programmers even? Looking to use it for a career change. Local uni grad with engineering background here.
 

b0ring

Senior Member
Joined
Aug 9, 2019
Messages
585
Reaction score
4
I saw there is a JCU online master of data science. You complete a number of module you get graduate certificate, then a few more become diploma, and a few more become master. But unsure how good the course is.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Bro,

I strongly recommend not taking the Data Science courses offered by any Poly in Singapore if you truly want to do a career change in Data Science.

But before I prescribe anything to you, can I just understand what is your background like? What did you do for your bachelor's (if any). This will allow me to understand how much math you have before I can prescribe something more suitable for you!
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Boring,

Haha I don't really think it is worth it. If we are talking about online masters, nothing can beat the Georgia Tech's OMCS in terms of value for money. Georgia Tech is one of the top CS schools in the world, and you only need pay 12k++ to complete the degree.

In terms of material, you go through the same course as the onsite students do. If you google around the data scientists in singapore who are working in Lazada, Grab etc, you will see that quite a few of them are enrolled in this program. So this program is definitely recognised.

But BIG disclaimer. This program is not easy. You can expect to spend at least 20 hours a week for lessons + assignments. This is if you intend to complete the program in 2 years. So it really requires careful planning, and you need to have some math/programming background.
 

Skyeagle

Senior Member
Joined
Jun 3, 2003
Messages
829
Reaction score
21
Hello DataScience,

Appreciate if you can provide some of your inputs:

1. Can a software engineer go to the data science field or some one who is more of a data science person go to the software engineering field?

2. Is it true that the FAANG companies pay a lot more in the US if you are working for them as a data scientist as compare to the other non-FAANG companies?

3. In your opinions, what are the expectations of a "Data Scientist"? Somehow it seems that the title data scientist is mis-used in local industries.

Thanks.
 
Important Forum Advisory Note
This forum is moderated by volunteer moderators who will react only to members' feedback on posts. Moderators are not employees or representatives of HWZ. Forum members and moderators are responsible for their own posts.

Please refer to our Community Guidelines and Standards, Terms of Service and Member T&Cs for more information.
Top