tHe237046
Badges: 10
Rep:
?
#1
Report Thread starter 2 months ago
#1
Hi

Im an aspiring data scientist studying computer engineering.
I currently work at a data and market research firm to support myself through my studies.

Im deeply interested in data science and quantitative finance. Mainly due to the mathematics/challenge/prestige and pay.
Ive good knowledge of basic mathematics and engineering mathematics, especially control engineering and signals and systems, Ive good grades in programming modules. Ive got good knowledge upto A level economics and some basic finance, but nothing quantitative.

Where do I start with data science? Im so confused by the number of courses online and information, its honestly overwhelming.

I am also interested in quantitative finance and algorithmic trading, a lot of which seems to overlap with data science?

I can do an internship in the firm that I work in, but how do I learn actual data science and quantitative analysis? Would appreciate advice from those whove been down this path.

Im not in the UK, so this post is mainly about where you learnt what you know for these two fields, and how can a beginner with near zero knowledge start.
Last edited by tHe237046; 2 months ago
0
reply
VannR
Badges: 19
Rep:
?
#2
Report 2 months ago
#2
I'm a recent Maths grad with a lot of programming and comp sci experience, and for someone with that background, the best learning pathway appears to be a mixture of Udemy courses and math-heavy textbooks (libgen is your
1
reply
tHe237046
Badges: 10
Rep:
?
#3
Report Thread starter 2 months ago
#3
(Original post by VannR)
I'm a recent Maths grad with a lot of programming and comp sci experience, and for someone with that background, the best learning pathway appears to be a mixture of Udemy courses and math-heavy textbooks (libgen is your
I have the programming and compsci experience, in the form of my own programs and projects. May I ask where you are working now?

Yeah I know libgen and have most of the mathematics books in real analysis, signals and control theory, abstract algebra, etc. I dont have the abstract maths knowledge but I know axioms,definitions, theorems and can do challenging mathematics and proofs (real analysis self studied it). But its a little overwhelming, I wish I could have gone for a mathematics degree though but its too late for that.
Last edited by tHe237046; 2 months ago
0
reply
tHe237046
Badges: 10
Rep:
?
#4
Report Thread starter 2 months ago
#4
(Original post by VannR)
I'm a recent Maths grad with a lot of programming and comp sci experience, and for someone with that background, the best learning pathway appears to be a mixture of Udemy courses and math-heavy textbooks (libgen is your
Ive no knowledge of partial differential equations, so the black scholes model and brownian motion stuff has been really over my head.

Were only taught LINEAR ODEs as we mostly deal with LTI systems in our course.

What do you think? Should I do a masters/PhD in signal processing? Or a masters in quantitative finance like an MFE?
Because I believe the data science is the safer route, but it all has to be self taught. However a lot of the job openings for quant roles and on quant websites ask for data scientists which, is something Im not really able to distinguish for these roles.
Last edited by tHe237046; 2 months ago
0
reply
VannR
Badges: 19
Rep:
?
#5
Report 2 months ago
#5
(Original post by tHe237046)
Ive no knowledge of partial differential equations, so the black scholes model and brownian motion stuff has been really over my head.

Were only taught LINEAR ODEs as we mostly deal with LTI systems in our course.

What do you think? Should I do a masters/PhD in signal processing? Or a masters in quantitative finance like an MFE?
Because I believe the data science is the safer route, but it all has to be self taught. However a lot of the job openings for quant roles and on quant websites ask for data scientists which, is something Im not really able to distinguish for these roles.
OK, there are a lot of questions you are asking here, but I will try to go through them methodically.

My undergraduate specialisms were statistics, probability theory, and applications of calculus to financial applications. Black-Scholes theory is bread-and-butter work for me. I took a year of computer science beforehand, did very well at it, but realised I preferred Mathematics and changed degrees. My academic background and experience with speaking to recruiters and applying for jobs has given me the following persepctive:

Quantitative finance is, traditionally, the field where methods in calculus, optimisation theory and probability theory are applied to asset pricing and portfolio analysis, as well as the creation and analysis of trading models. In recent times, the power of machine learning algorithms means that much of this analysis can be "black-boxed" i.e. it can be passed to a system such as a neural network which will produce effective results, but will not give semantics about its operation. This has pros and cons, and in fact it is often the case that firms are not overly interested in "data scientists" who do not know anything more than how to apply "black box" methods.

However, data science in general is a much broader technical skillset to have. You need to understand data capture, warehousing and cleaning before you even begin to analyse a dataset using the specific modern tools of machine learning (clustering, regression, SVMs etc), and it is applicable to almost every single major industry in the modern world. In a data science job, you will be expected to be able to apply your technical know-how in a domain-specific manner, in order to produce actionable insights that your colleagues can use in order to improve the business.

Quantitative finance is a mathematical discipline with a very particular use of the techniques of data science. Data science is far more general, use-case focused, and its mathematical side is usually only important for high-end researchers, not practitioners.

For someone with your academic background, I would highly advise that you look at data engineering as well as data science. This is the particular side of the data industry where people with specific knowledge of systems and data manipulation shine. In essence, data engineers are librarians. They figure out how data needs to be captured, systematised, and stored. Data scientists are then people who take out library books. They take the data and apply their methods to produce insights. There is often a great overlap in these roles, and both are respected and can lead to great salaries.

In your position, if you cannot self-teach or produce a good portfolio of data science projects on your own, take an MSc in Data Science. An MFE is a really niche (and expensive) thing to pursue, and unfortunately you will find that your lack of Mathematics knowledge will hold you back if you want to be a quant without significant self-study into what you lack.
0
reply
tHe237046
Badges: 10
Rep:
?
#6
Report Thread starter 2 months ago
#6
(Original post by VannR)
OK, there are a lot of questions you are asking here, but I will try to go through them methodically.

My undergraduate specialisms were statistics, probability theory, and applications of calculus to financial applications. Black-Scholes theory is bread-and-butter work for me. I took a year of computer science beforehand, did very well at it, but realised I preferred Mathematics and changed degrees. My academic background and experience with speaking to recruiters and applying for jobs has given me the following persepctive:

Quantitative finance is, traditionally, the field where methods in calculus, optimisation theory and probability theory are applied to asset pricing and portfolio analysis, as well as the creation and analysis of trading models. In recent times, the power of machine learning algorithms means that much of this analysis can be "black-boxed" i.e. it can be passed to a system such as a neural network which will produce effective results, but will not give semantics about its operation. This has pros and cons, and in fact it is often the case that firms are not overly interested in "data scientists" who do not know anything more than how to apply "black box" methods.

However, data science in general is a much broader technical skillset to have. You need to understand data capture, warehousing and cleaning before you even begin to analyse a dataset using the specific modern tools of machine learning (clustering, regression, SVMs etc), and it is applicable to almost every single major industry in the modern world. In a data science job, you will be expected to be able to apply your technical know-how in a domain-specific manner, in order to produce actionable insights that your colleagues can use in order to improve the business.

Quantitative finance is a mathematical discipline with a very particular use of the techniques of data science. Data science is far more general, use-case focused, and its mathematical side is usually only important for high-end researchers, not practitioners.

For someone with your academic background, I would highly advise that you look at data engineering as well as data science. This is the particular side of the data industry where people with specific knowledge of systems and data manipulation shine. In essence, data engineers are librarians. They figure out how data needs to be captured, systematised, and stored. Data scientists are then people who take out library books. They take the data and apply their methods to produce insights. There is often a great overlap in these roles, and both are respected and can lead to great salaries.

In your position, if you cannot self-teach or produce a good portfolio of data science projects on your own, take an MSc in Data Science. An MFE is a really niche (and expensive) thing to pursue, and unfortunately you will find that your lack of Mathematics knowledge will hold you back if you want to be a quant without significant self-study into what you lack.
Ok.

Theres some learning I could do for quantitative analysis, but definitely it cannot stack up against the smattering of courses you have taken, despite me being a self taught genius, my time is limited due to actual work, study and life commitments.

Ill read around some more and then make up my mind at the end of this month. It seems quant is out without me doing a Phd Or masters, or going for another bachelors in mathematics. These are all crazy options and not really realistic. Ive been reading the introductary finance book and that is painful as well as Paul Willmotts. Data science definitely seems more achievable as well as seems to have more jobs and growth.

My last query:

Can someone not having taken such in depth courses in statistics and mathematics at university, self teach himself enough to be taken on as a quant developer or even a quant analyst, if he teaches himself through books like Pual Willmotts introduces quantitative finance?

I have the programming background for both domains, and general understanding of mathematical philosophy and framework to launch myself without a sweat. I just dont have the courses you pure math guys take, so do you think niche books like Paul Willmotts intro and C++ in algo trading, are enough for becoming a quant developer - A glorified programmer but well-paid and easy to find a job?
What sorts of quants are there?
(1) Front office/desk quant
(2) Model validating quant
(3) Research quant
(4) Quant developer

Seems number 4 is the easiest for computer science and engineering graduates.
Also I see you are doing a PhD in math.

Im very sorry to be so obnoxious, I have not many places to ask for serious advice. Also its something I only started researching slowly over the past 2 months and not really commmitted to deeply, so my apologies for the obnoxiousness.
Your answer was really quite clear and detailed, thank you!!
Last edited by tHe237046; 2 months ago
0
reply
VannR
Badges: 19
Rep:
?
#7
Report 1 month ago
#7
(Original post by tHe237046)
Ok.

Theres some learning I could do for quantitative analysis, but definitely it cannot stack up against the smattering of courses you have taken, despite me being a self taught genius, my time is limited due to actual work, study and life commitments.

Ill read around some more and then make up my mind at the end of this month. It seems quant is out without me doing a Phd Or masters, or going for another bachelors in mathematics. These are all crazy options and not really realistic. Ive been reading the introductary finance book and that is painful as well as Paul Willmotts. Data science definitely seems more achievable as well as seems to have more jobs and growth.

My last query:

Can someone not having taken such in depth courses in statistics and mathematics at university, self teach himself enough to be taken on as a quant developer or even a quant analyst, if he teaches himself through books like Pual Willmotts introduces quantitative finance?

I have the programming background for both domains, and general understanding of mathematical philosophy and framework to launch myself without a sweat. I just dont have the courses you pure math guys take, so do you think niche books like Paul Willmotts intro and C++ in algo trading, are enough for becoming a quant developer - A glorified programmer but well-paid and easy to find a job?
What sorts of quants are there?
(1) Front office/desk quant
(2) Model validating quant
(3) Research quant
(4) Quant developer

Seems number 4 is the easiest for computer science and engineering graduates.
Also I see you are doing a PhD in math.

Im very sorry to be so obnoxious, I have not many places to ask for serious advice. Also its something I only started researching slowly over the past 2 months and not really commmitted to deeply, so my apologies for the obnoxiousness.
Your answer was really quite clear and detailed, thank you!!
It is great that you have a good understanding of the different roles that can call themselves "quants". In terms of function, you can think of them as researchers or developers. Researchers have an MSc, of often a Ph.D in statistics/PDEs/machine learning. Their jobs are to create (as in actually invent) financial products which they can justify as appropriate for the current purposes using advanced mathematics. This job is highly mathematical, incredibly fast-paced, and the pay can become incredible. We're talking £200k base salary with up to 100% bonuses for performance for people who are good. Quantitative developers implement what these people theorise. They need to understand computational methods, low-latency computing, and evaluation of trading methods. These people may not command the same salary as the researchers, but importantly, there is a greater supply and more of them are often required in a particular institution than researchers. You can earn £100k+ a year, but you may not be seen as a "revenue generator", and hence won't get such generous bounses.

Paul Wilmott's books are absolutely fantastic for someone who needs to understand the theory of financial products and the basics of methods used to evaluate them, and the different tradeable assets which exist - one can trade volatility itself!

Having a look at your Quora, it is clear that you are motivated and that you are looking to do whatever it takes. If you are really in your heart looking for a quantitative researcher/developer hybrid role, you need to learn the following mathematics:

- Linear Algebra (theoretical, focus on understanding spaces and the link to Sturm-Liouville Theory)
- Real Analysis (differentiation, Riemann integrals, function spaces, metric spaces)
- Probability Theory (up to Time Series Analysis, renewal processes, Markov chains)
- Stochastic Calculus (up to Ito's Formula, derivation of BS-equation, pricing of Asian average options and bonds)
- Calculus (multivariate calculus, vector analysis, Sturm-Liouville Theory, PDE methods(analytic and numerical), Fourier Analysis)

This is what underpins quantitative analysis. If you want to break into the calculus of variations and the Malliavian calculus that is the true handbook of quantitative researchers, you cannot escape this knowledge. You can build the knowledge over time, but you have to actually do problems in order to gain any proficiency.

As for myself, financial limitations have caused me to postpone my Ph.D for a year, but I am working on whatever I can in the mean time.
1
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

Are you registered to vote?

18-20 years old (yes) (135)
54%
18-20 years old (no) (30)
12%
20-25 years old (yes) (39)
15.6%
20-25 years old (no) (4)
1.6%
25-30 years old (yes) (18)
7.2%
25-30 years old (no) (0)
0%
30-40 years old (yes) (12)
4.8%
30-40 years old (no) (1)
0.4%
40+ years old (yes) (5)
2%
40+ years old (no) (6)
2.4%

Watched Threads

View All