Advice on BSc Comp Sci and Mathematics

[Siyachuma]

Hello
My main interest is in analytics, data science and business, with a little bit of online research it seems a good starting point would be to study a BSc on Computer science and Mathematics, but I'm not really sure.

So I would like to find out from someone who has done such courses what their advice is. I'm a newly qualified pharmacist at the moment and would like to learn something else on the side over the next couple of years and as such I'm inclined to consider Unisa

Practical advice would be highly appreciated

 

[cguy]

Definitely a good path. Make sure to throw in as much stats and applied maths as possible with it. This will be a bit more relevant to your goals than pure maths.

I also suggest going at least up to honours with this.

 

[Siyachuma]

He

Definitely a good path. Make sure to throw in as much stats and applied maths as possible with it. This will be a bit more relevant to your goals than pure maths.

I also suggest going at least up to honours with this.

Thanks for your response, so wouldn't a different stream be better choice for me.

 

[cguy]

He
Thanks for your response, so wouldn't a different stream be better choice for me.

You probably don’t have to choose your final stream at this point. I did a BSc CS + Maths myself, but I managed to get in 1 year of stats and 2 years of applied maths as well. If I was doing this again, I probably would have tried to squeeze in Stats 2 even for non-degree purposes, and then depending on the modules perhaps done more app maths and stats to 3rd year too.

As it turns out, topology, complex analysis, etc. aren’t terribly useful for data analysis. . They did however give me a good mathematical background, which helped a lot studying further and career wise.

Make sure you do well in the CS courses - at the end of the day, this is what you will be using to actually “do” anything with your maths skills.

 

[Siyachuma]

T

You probably don’t have to choose your final stream at this point. I did a BSc CS + Maths myself, but I managed to get in 1 year of stats and 2 years of applied maths as well. If I was doing this again, I probably would have tried to squeeze in Stats 2 even for non-degree purposes, and then depending on the modules perhaps done more app maths and stats to 3rd year too.

As it turns out, topology, complex analysis, etc. aren’t terribly useful for data analysis. . They did however give me a good mathematical background, which helped a lot studying further and career wise.

Make sure you do well in the CS courses - at the end of the day, this is what you will be using to actually “do” anything with your maths skills.

Hank
Omw, I'm looking at the at the curriculum at Unisa and thinking it's jam-packed as is and you're saying it would be better to add more subjects into the mix. I hope this makes sense as I'm not familiar with the structure of BSc degrees.
I'm also clueless as to which subject are the concepts you mentioned. I see you say Computer Science is the most important or fundamental subject for data science and then what's the best subject to pair it with between math, applied math and statistics. Sorry if I'm asking you to repeat yourself, perhaps your explanation went a bit above my head

 

[cguy]

T
Hank
Omw, I'm looking at the at the curriculum at Unisa and thinking it's jam-packed as is and you're saying it would be better to add more subjects into the mix. I hope this makes sense as I'm not familiar with the structure of BSc degrees.
I'm also clueless as to which subject are the concepts you mentioned. I see you say Computer Science is the most important or fundamental subject for data science and then what's the best subject to pair it with between math, applied math and statistics. Sorry if I'm asking you to repeat yourself, perhaps your explanation went a bit above my head

So, CS will help with what you actually end up doing. You will likely eventually be analyzing data in R or Python (currently trending in this area), or if you actually end up working on the design and implementation of the algorithms themselves, C++. To do this robustly, and at scale a good CS background is useful.

Many of the machine learning algorithms draw from a combination of applied maths and statistics. The applied maths (specifically calculus based parts) are important for understanding the various machine learning algorithms that attempt to minimize error.

Since you typically deal with many variables, most of the algorithms are built on top of matrix algebra, so “linear algebra” (usually 2nd year pure maths), is a really good base for this.

Statistics is important for understanding the properties of your actual data and nature of your variables, predictions, errors, etc.

You don’t have to cover them all in your degree, that is more of just a best case, but it may be too much work. When you get enough of a grasp on the fundamentals you can fill in the gaps yourself.

Key quantitative coursework:
- Maths: calculus, linear algebra
- Applied Maths: differential equations, numerical methods, applied calculus
- Statistics: probability, distributions, linear regression

 

[Siyachuma]

So, CS will help with what you actually end up doing. You will likely eventually be analyzing data in R or Python (currently trending in this area), or if you actually end up working on the design and implementation of the algorithms themselves, C++. To do this robustly, and at scale a good CS background is useful.

Many of the machine learning algorithms draw from a combination of applied maths and statistics. The applied maths (specifically calculus based parts) are important for understanding the various machine learning algorithms that attempt to minimize error.

Since you typically deal with many variables, most of the algorithms are built on top of matrix algebra, so “linear algebra” (usually 2nd year pure maths), is a really good base for this.

Statistics is important for understanding the properties of your actual data and nature of your variables, predictions, errors, etc.

You don’t have to cover them all in your degree, that is more of just a best case, but it may be too much work. When you get enough of a grasp on the fundamentals you can fill in the gaps yourself.

Key quantitative coursework:
- Maths: calculus, linear algebra
- Applied Maths: differential equations, numerical methods, applied calculus
- Statistics: probability, distributions, linear regression

Thanks for your information and advice

 

[bchip]

Hello
My main interest is in analytics, data science and business, with a little bit of online research it seems a good starting point would be to study a BSc on Computer science and Mathematics, but I'm not really sure.

Programming and data analytics are 2 different career paths in today's world.
The 1 mentioned above is the programming + mathematics, another path is Informatics

Data scientists have a mathematical background
Data analysts have an informatics background.

Data analysts and data architects look at how to setup a database store the information
and then use it to make more general reports.
Data scientists are more specialized mathematicians that use math to do certain tests.

 

[cguy]

Programming and data analytics are 2 different career paths in today's world.
The 1 mentioned above is the programming + mathematics, another path is Informatics

Data scientists have a mathematical background
Data analysts have an informatics background.

Data analysts and data architects look at how to setup a database store the information
and then use it to make more general reports.
Data scientists are more specialized mathematicians that use math to do certain tests.

^ yup. The last time I defined a database schema was 1997.

 

[Siyachuma]

Programming and data analytics are 2 different career paths in today's world.
The 1 mentioned above is the programming + mathematics, another path is Informatics

Data scientists have a mathematical background
Data analysts have an informatics background.

Data analysts and data architects look at how to setup a database store the information
and then use it to make more general reports.
Data scientists are more specialized mathematicians that use math to do certain tests.

Hello
Thanks for your response. I am not getting the practical consequence of your input. My interest is data science, my understanding of that is using data to make 'smart' decisions. It seems the processes of a data scientist are based on statistics and applied mathematics, if I understand the previous advice correctly. However, mathematics and Computer science is a good foundation to learn these skills in the future as one cannot learn everything at once or in one degree program.
Another aspect, please correct me if I am wrong, practically one would be a good scientist if they start of as an analyst. Are you then saying this would require me to learn or prioritize informatics. Please clarify

 

[cguy]

Hello
Thanks for your response. I am not getting the practical consequence of your input. My interest is data science, my understanding of that is using data to make 'smart' decisions. It seems the processes of a data scientist are based on statistics and applied mathematics, if I understand the previous advice correctly. However, mathematics and Computer science is a good foundation to learn these skills in the future as one cannot learn everything at once or in one degree program.
Another aspect, please correct me if I am wrong, practically one would be a good scientist if they start of as an analyst. Are you then saying this would require me to learn or prioritize informatics. Please clarify

The data science side tends to focus more on advanced processing of mostly opaque data. The data analytics side tends to be less mathematically complex, but has a larger focus on context and interpretation.

I tend to do a bit of both: Predicting market prices for trading is very much a data science problem.

Doing post-trade analysis to see how well the strategies are doing, determining in what ways the results don’t match expectations, and making decisions on how to rectify them is very much a data analytics problem.

 

[bchip]

Hello
Thanks for your response. I am not getting the practical consequence of your input. My interest is data science, my understanding of that is using data to make 'smart' decisions. It seems the processes of a data scientist are based on statistics and applied mathematics, if I understand the previous advice correctly
...

Sorry I should have been more clear. The areas that you mentioned have very much different career paths
and I dont believe that people are always aware of the different ones.
So I just wanted to point out that mathematics and programming is not the only route into that world.
If your looking at the specifically data scientist, you enjoy math then as cguy mentioned
BSc CS, BSc Applied Mathematics, BSc/BCom statisitcs, BCom Actuarial sciences...all of these will get you there.

If mathematics or programming is not your strong suit but you are still interested in the data world another
option would be BCom Informatics (just that you are aware of it).

Sounds like you know which direction you are aiming for.
There are a couple of courses on coursera and edx (hosted by Microsoft) that are only a few weeks
long and can be done without paying that you can also try out - if you wanted to see what its like
before committing to a full time course.

Best of luck

 

[Siyachuma]

Sorry I should have been more clear. The areas that you mentioned have very much different career paths
and I dont believe that people are always aware of the different ones.
So I just wanted to point out that mathematics and programming is not the only route into that world.
If your looking at the specifically data scientist, you enjoy math then as cguy mentioned
BSc CS, BSc Applied Mathematics, BSc/BCom statisitcs, BCom Actuarial sciences...all of these will get you there.

If mathematics or programming is not your strong suit but you are still interested in the data world another
option would be BCom Informatics (just that you are aware of it).

Sounds like you know which direction you are aiming for.
There are a couple of courses on coursera and edx (hosted by Microsoft) that are only a few weeks
long and can be done without paying that you can also try out - if you wanted to see what its like
before committing to a full time course.

Best of luck

Thank you, I will do further reading into the courses and see if I'll finally do that application for this particular degree or perhaps another one of the ones you've mentioned. Thanks for the valuable information.

 

[cheesus]

To be quite frank before I employ a University to teach me anything to do with technology I would much rather get either a Lynda.com / Udemy or Pluralsight account. Cost a couple of hundred bucks a month and will get you literate in the technologies you are interested much faster.

 

[cguy]

To be quite frank before I employ a University to teach me anything to do with technology I would much rather get either a Lynda.com / Udemy or Pluralsight account. Cost a couple of hundred bucks a month and will get you literate in the technologies you are interested much faster.

This is a terrible idea for anything involving data science or mathematics.

 

[cheesus]

This is a terrible idea for anything involving data science or mathematics.

why?

 

[cguy]

why?

The tech stuff one can watch videos on, read books on and get good experience in by practicing at home to get hands on experience.

Maths, stats, etc. is far less intuitive and has far less self evident application, or at least means to exercise it. It is also a highly dependent tree of study, where missing out on some course you’ve never heard of, or not fully internalizing something that didn’t seem important is likely to make studying the next thing you want to study nearly impossible.

Put another way, I know plenty of good self taught programmers, but almost zero good self taught mathematicians.

 

[Siyachuma]

You probably don’t have to choose your final stream at this point. I did a BSc CS + Maths myself, but I managed to get in 1 year of stats and 2 years of applied maths as well. If I was doing this again, I probably would have tried to squeeze in Stats 2 even for non-degree purposes, and then depending on the modules perhaps done more app maths and stats to 3rd year too.

As it turns out, topology, complex analysis, etc. aren’t terribly useful for data analysis. . They did however give me a good mathematical background, which helped a lot studying further and career wise.

Make sure you do well in the CS courses - at the end of the day, this is what you will be using to actually “do” anything with your maths skills.

Say I choose this particular bsc in comp Sci and math at unisa . Does it contain enough mathematics to gain entry into a comp Sci and applied math honours ( at least a second year in a mathematics equivalent).

 

[Knyro]

Say I choose this particular bsc in comp Sci and math at unisa . Does it contain enough mathematics to gain entry into a comp Sci and applied math honours ( at least a second year in a mathematics equivalent).

You might want to consider a general BSc as well at Unisa instead of a particular stream. This way you can pick ALL the modules in your degree. Just make sure you include the necessary third year modules for the majors you want then work backwards, filling in all pre and co-requisites as you go. Then you can fill in any space with modules from something else.

For example I did a physics and CS general BSc, took enough physics and CS modules to get into honours for both , but I took many maths & applied maths modules as well

 

[cguy]

Say I choose this particular bsc in comp Sci and math at unisa . Does it contain enough mathematics to gain entry into a comp Sci and applied math honours ( at least a second year in a mathematics equivalent).

For CS, yes - any major in a subject should be sufficient for honours if your marks are high enough.

So for maths, you could do maths honours, but not necessarily App Maths honours. This would be up to the department’s discretion, and will depend on how well you did in CS and Maths, which subjects you took and what honours modules they offer. If most honours modules rely heavily on App Maths 3, you probably won’t be allowed to do it unless they think you can self study what you are missing.