Imperial MSc Machine Learning and Data Science (Online via Coursera)
Scroll to see replies
Original post by basedlines
It's tough to say. It would depend on your background, your value hierarchy and what your aims are in undertaking the degree (career? academia? 4fun?).I'm definitely enjoying the course, but I wouldn't be able to say if the delivery of content is exactly 2x more compelling/engaging/informative than would be available on other comparable MSc courses (assuming other typical MScs are priced around £15k). How much you are willing to pay for this particular degree over others will depend on how highly you value other factors, such as the Imperial brand, networking with coursemates and the flexibility of a part time degree over a full time degree.More generally, if you aren't yet sure whether to pursue an MSc in this space and you're just aiming to gain minimally sufficient competence for a career transition into a technical role like ML engineering, then most of the practical skills taught on the course could realistically be learnt over a similar timeframe without the £30k expense through committed self study and independent project work. (Although the accountability of pursuing the degree helps to keep that goal close to heart). The real benefit to me thus far has been in gaining a robust understanding of the statistics and mathematics, to a significantlly deeper level of rigour than would have been achievable through independent study.Please take all of this with a grain of salt since only 1 semester out of 6 total has been delivered thus far, and there are many more modules later in the course which seem to have a technical focus.
Thanks so much for the info! I'm also interested in applying for the course and your comments have been so insightful, thanks
I also have quite a few questions to ask:
- I know you're only in your 1st term but I was just wondering if you know what kind of projects are offered for the final research project/dissertation? will you have a supervisor like you did in undergrad? can you contact them/work collaboratively?
- How would you rate the communication between students/professors/course coordinators? Can you ask questions if you have any? Is the student support comparable to that of a normal degree?
- How would you rate your overall satisfaction with the course?
- Did you consider any other unis/msc's when applying for this?
- How are the assessments structured - e.g. is there a specific date for each assessment/what if you're busy with work?
- How flexible is the course? and how is your work/life balance in terms of handling this and a full-time job? Do you think it's manageable? I'm going to start a new job in October, so I'm not really sure if it'll be the best time to start a part-time MSc at the same time, but I'm also interested in this course because I just really love learning/studying in my free time and I feel like it'll give me a structured & more robust experience to learn what I need to (as opposed to self-learning/learning through a job) in order to get into a more specialized data science role!
- Lastly, I don't have an extensive mathsy background (I was in biomed, but have been working as a data analyst for about a year now and have been loving it + I learnt a lot of coding/basic stats through the job/self-learning) - is there anything you recommend in order for me to sell myself better in the application + prepare myself better for the course?
(edited 2 years ago)
Original post by chickenpotpiie
Thanks so much for the info!
There hasn't been much discussion yet about particular dissertation project titles, but I'll let you know soon as.
The student support is actually a strong point for the course. We have at least two live sessions per module per week with one live lecture and one open office hour. These live sessions give you a good opportunity to ask any questions. On top of this we have two online forums. One is a weekly forum on coursera which is intended to be used for questions relating to weekly content. The other forum is on edstem.org, with boards for each module where you can ask more general questions about ideas/themes/topics etc. Broader discussion and curiosity are definitely encouraged. This week a couple people posted passages from relevant textbooks/papers that they've read, and the lecturers responded to questions about them.
Overall I'm satisfied. The course price is definitely a bit inflated but if I could go back I reckon I'd make the same choice.
I had offers for these MSc courses:
• Nottingham - Machine Learning in Science,
• Nottingham - Computer Science (AI) [2Years Full Time],
• Edinburgh - Data Science, Technology and Innovation,
• UCL - Data Science and Machine Learning [Full Time],
• Bath - Artificial Intelligence.
I was considering the Nottingham CompSci course, because in the second year you produce extended research which will hopefully become published work. At the time I was considering a PhD and the extended research seemed like good experience. In the end I went for Imperial mostly because of the prestige and the expectation of better teaching quality.
There are no timed exams. All assessments come in the form of small projects/courseworks. Typically you will have around 5-10 days (depending on size of project) to complete the coursework at your own pace. This period will usually fall over a weekend or two, to help make sure you have sufficient time to complete the tasks. Each project will account for some % of the final mark for each module, with the final projects accounting for a large majority. It's your responsibility to plan your time so that you can complete the project by the due date. Unfortunately there's very little flexibility to ask for extensions without good reason as far as I can tell.
I reckon you'd manage. The majority of folks on the course are working full time or part time jobs as well. It is definitely a demanding course, but if you stay committed and put aside time in the evenings and weekends then it's manageable. Since you're working as a data analyst, most of what you learn will likely feel directly relevant to your work so it won't be a big mental jump to swap back and forth between work and studying. You do have to be mindful that you won't have a lot of time for other hobbies and committments. Luckily we do have Christmas, Spring and Summer break like normal students.
Definitely mention your data analyst work and highlight key technical skills you've learnt. To be honest I would bet money that demonstrating competence in maths is a huge factor for the admissions team. Make sure you mention any formal qualifications like maths a-level or maths modules from your undergrad to show that you have an affinity for maths. Mention that you've been doing wider reading in your own time to learn key topics (especially Calculus, Linear Algebra, Statistics, Probability). For example:
• "Intro to Probability for Data Science" (available for free at probability4datascience.com)
• "Mathematics for Machine Learning",
• "An Introduction to Statistical Learning with application in R", <-- really great!
• the first few chapters of "Deep Learning" by Ian Goodfellow.
If those feel a bit inaccessible then I'd recommend "Bridging the Gap to University Maths", which covers a lot of fundamental basics in an unassuming way.
(edited 2 years ago)
Original post by slentoris
Bad experience as a student in the current cohort. Don't sign up. Waste of money and time.
That's interesting... why do you say that?
Original post by slentoris
Bad experience as a student in the current cohort. Don't sign up. Waste of money and time.
Where are you studying? pl. give insights on how it is bad
Original post by basedlines
The course is demanding. This semester I spent roughly 30 hours per week studying....
Thanks basedlines - your posts on here have been really useful.
You were saying that a lot of your fellow students had full time jobs. That's impressive if they're managing that with an additional 30 hours for studying! Or are their employers giving them study leave for it?
Now you've done another few terms, would you say the hours have stayed about the same per week? Do the students with full time jobs seem to be keeping up or is there a noticeable difference compared to those with more available time?
30 hours is equivalent to four days a week study leave, which seems like a lot to ask from an employer for 30 weeks a year over two years (albeit I guess that would be reduced by studying in evenings and weekends).
Also, is there any way that you can indicate the complexity of the course content? For example, some recent material or assessment questions? (Without sharing anything that you aren't allowed to, of course!)
(edited 2 years ago)
Original post by slentoris
Bad experience as a student in the current cohort. Don't sign up. Waste of money and time.
Can you elaborate?
In general the courses Big Data and Bayesian Methods have been excellent.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
Thank you for the information! Can you confirm that this masters is really focusing on the mathematics behind all the machine learning algorithms (which are very popular now and almost everywhere when teaching they put attention mainly on the python functions and the programming…)? I chose Imperial because of this claim as a person who graduated in Applied Mathematics.
Original post by basedlines
In general the courses Big Data and Bayesian Methods have been excellent.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
The course is taught by the Maths/Stats faculty, so there is definitely an appreciation and focus on maths.
The Supervised Learning module essentially taught out of the book "Introduction to Statistical Learning" by Hastie and Tibshirani.
Take a look at that book to get a good idea of what you will learn/know.
pdf available for free at: https://hastie.su.domains/ISLR2/ISLRv2_website.pdf
The Supervised Learning module essentially taught out of the book "Introduction to Statistical Learning" by Hastie and Tibshirani.
Take a look at that book to get a good idea of what you will learn/know.
pdf available for free at: https://hastie.su.domains/ISLR2/ISLRv2_website.pdf
(edited 2 years ago)
Original post by basedlines
In general the courses Big Data and Bayesian Methods have been excellent.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
These two courses have been rigorous, with high quality notes and well produced video content.
Very high educational value.
Here is a copy of my most recent Bayesian coursework submission.
We are expected to complete courseworks of approximately this length every week, usually containing a mix of code and math.
It's been quite tight on time, many have swapped to part time work, and a handful have dropped out.
Of the people I see staying full time, many are working already as data scientists.
I personally wouldn't recommend working full time >40 hours a week on top of the course unless you are especially committed.
There have been a few teething issues for the course, most of which have been resolved.
The consensus among our cohort is that some modules, like the Programming for Data Science module, have been thoroughly underwhelming.
This is likely why the other poster has commented "a waste of time and money".
Perhaps by next year the issues will be fixed entirely.
In any case I'm happy with the course and plan to continue on to the second year.
After your recommendations, I applied and am now in the Oct 2022 - Oct 2024 cohort!
They must have taken feedback from previous years into account as the Programming for Data Science module this term is feeling very useful so far.
Hopefully your second year is going really well. It will be great to hear your experience of the dissertation as I'm really looking forward to that.
Original post by apples_galois
After your recommendations, I applied and am now in the Oct 2022 - Oct 2024 cohort!
They must have taken feedback from previous years into account as the Programming for Data Science module this term is feeling very useful so far.
Hopefully your second year is going really well. It will be great to hear your experience of the dissertation as I'm really looking forward to that.
They must have taken feedback from previous years into account as the Programming for Data Science module this term is feeling very useful so far.
Hopefully your second year is going really well. It will be great to hear your experience of the dissertation as I'm really looking forward to that.
sounds great and happy for you! Would you be more specific about how it is useful? cheers
Original post by JackoC
sounds great and happy for you! Would you be more specific about how it is useful? cheers
Sure! There are lots of ways in which I find this particular module helpful.
I appreciate the fact we learn both Python and R. I was self-taught in Python and had no R experience before the course, so it's great to have an additional language. I still prefer Python, but I think thoroughly learning multiple languages is really helpful for understanding deeper programming concepts.
The lecturers are always happy to answer questions - even those that are only tangentially related to the course content! - and their live sessions are always thought-provoking.
There's a real spectrum of ways of learning - hands-on coding (both exercises and guided 'worked examples'), textbooks, videos, live sessions, questions, academic papers etc. I'm sure that everyone has a preferred method of learning, but I don't think there's "one best method" and there's always going to be something more you gain from a different source/style.
It's also a more broad course than "how to code in Python and R". It includes content that is more practical if you're writing software day-to-day, as well as more theoretical (e.g. language-agnostic coding best-practice).
Then there's the sort of benefits you get from any course. The fact that you have deadlines and a course structure to follow is a great source of motivation. One can buy textbooks or watch the many free tutorials or pay for access to an online learning platform, but it is a cliche how quickly that initial enthusiasm wanes for self-taught stuff. It can easily be added to the large pile of unfinished projects. But a formal course like this keeps you on track.
Likewise, you get a cohort of around 50 students with whom you can discuss content, ask questions and share resources.
I'm sure I could think of lots of other good things about this module, but these are the ones that sprung to mind.
Original post by apples_galois
Sure! There are lots of ways in which I find this particular module helpful.
I appreciate the fact we learn both Python and R. I was self-taught in Python and had no R experience before the course, so it's great to have an additional language. I still prefer Python, but I think thoroughly learning multiple languages is really helpful for understanding deeper programming concepts.
The lecturers are always happy to answer questions - even those that are only tangentially related to the course content! - and their live sessions are always thought-provoking.
There's a real spectrum of ways of learning - hands-on coding (both exercises and guided 'worked examples'), textbooks, videos, live sessions, questions, academic papers etc. I'm sure that everyone has a preferred method of learning, but I don't think there's "one best method" and there's always going to be something more you gain from a different source/style.
It's also a more broad course than "how to code in Python and R". It includes content that is more practical if you're writing software day-to-day, as well as more theoretical (e.g. language-agnostic coding best-practice).
Then there's the sort of benefits you get from any course. The fact that you have deadlines and a course structure to follow is a great source of motivation. One can buy textbooks or watch the many free tutorials or pay for access to an online learning platform, but it is a cliche how quickly that initial enthusiasm wanes for self-taught stuff. It can easily be added to the large pile of unfinished projects. But a formal course like this keeps you on track.
Likewise, you get a cohort of around 50 students with whom you can discuss content, ask questions and share resources.
I'm sure I could think of lots of other good things about this module, but these are the ones that sprung to mind.
I appreciate the fact we learn both Python and R. I was self-taught in Python and had no R experience before the course, so it's great to have an additional language. I still prefer Python, but I think thoroughly learning multiple languages is really helpful for understanding deeper programming concepts.
The lecturers are always happy to answer questions - even those that are only tangentially related to the course content! - and their live sessions are always thought-provoking.
There's a real spectrum of ways of learning - hands-on coding (both exercises and guided 'worked examples'), textbooks, videos, live sessions, questions, academic papers etc. I'm sure that everyone has a preferred method of learning, but I don't think there's "one best method" and there's always going to be something more you gain from a different source/style.
It's also a more broad course than "how to code in Python and R". It includes content that is more practical if you're writing software day-to-day, as well as more theoretical (e.g. language-agnostic coding best-practice).
Then there's the sort of benefits you get from any course. The fact that you have deadlines and a course structure to follow is a great source of motivation. One can buy textbooks or watch the many free tutorials or pay for access to an online learning platform, but it is a cliche how quickly that initial enthusiasm wanes for self-taught stuff. It can easily be added to the large pile of unfinished projects. But a formal course like this keeps you on track.
Likewise, you get a cohort of around 50 students with whom you can discuss content, ask questions and share resources.
I'm sure I could think of lots of other good things about this module, but these are the ones that sprung to mind.
really really helpful!!! that's very kind of you! I greatly appreciate it!
forgive me as I have one more question.
I am sure the live session or online office hours make this online programme outstanding. But, for students from different time zone or continents, are live sessions also accessible?
Many thanks!
Original post by JackoC
really really helpful!!! that's very kind of you! I greatly appreciate it!
forgive me as I have one more question.
I am sure the live session or online office hours make this online programme outstanding. But, for students from different time zone or continents, are live sessions also accessible?
Many thanks!
forgive me as I have one more question.
I am sure the live session or online office hours make this online programme outstanding. But, for students from different time zone or continents, are live sessions also accessible?
Many thanks!
Live sessions are typically on a week day at around 9am or 5pm UK time (give or take an hour or two). Each live session will alternate between these times week to week, to try and be as accessible as possible.
Original post by JackoC
really really helpful!!! that's very kind of you! I greatly appreciate it!
forgive me as I have one more question.
I am sure the live session or online office hours make this online programme outstanding. But, for students from different time zone or continents, are live sessions also accessible?
Many thanks!
forgive me as I have one more question.
I am sure the live session or online office hours make this online programme outstanding. But, for students from different time zone or continents, are live sessions also accessible?
Many thanks!
I can't give any direct experience as I'm based in the UK, but they try to alternate the timing of any live content between the morning and afternoon UK time to maximise the amount which will be at a sensible time overseas. However, the sessions are all recorded so you can watch them back. There's also a question board, so if you had a question in advance that you wanted answered you could put it on the board and then watch the recording back to hear the answer (if they don't type a response on the board itself).
Hope that helps.
Quick Reply
Related discussions
- University of Leeds Vs. Queen Mary University of London
- MSc Statistics online/distance learning
- Online Masters in Computer Science - University of Keele
- Imperial MSc Environmental Data Science and Machine Learning interview
- University of Leeds Online Msc in artificial Intellegence
- Chemistry undergraduate apply for machine learning MSc
- Online MSc Computer Science with Data Analytics at University of York
- University Selection
- Data Science Uni
- Online Msc in Computer Science at Hertfordshire
- Official Imperial Postgraduate 2023 Applicants Thread
- Got 9s in maths and further maths a year early but not sure what to do now
- Online MSc Computer Science w Artificial Intelligence - Uni of York
- Application for UCL MSc Machine Learning
- Online Data Science MSc - University of London, Goldsmiths
- Do I have a good chances for imperial cs AI course?
- Online MSc Artificial Intelligence - University of Bath
- Choosing between Imperial and Cambridge for Data Science
- MSc recommendations for Computing?
- MSc Computer Science (conversion course) 2024-2025
Latest
Trending
Last reply 10 hours ago
MPhil in Politics and International Studies Cambridge Applicants Thread 2025Last reply 10 hours ago
University of Edinburgh 2025-26 Postgraduate ApplicantsLast reply 5 days ago
What does RCA ‘application closed’ mean please?Last reply 2 weeks ago
MSc Theoretical and Computational Chemistry - OxfordLast reply 2 weeks ago
MPhil in Scientific Computing 2024 - University of CambridgeLast reply 1 month ago
MPhil in Basic and Translational Neuroscience at Uni of Cambridge
