Data Science Landscape for Finance and Banking
Vložit
- čas přidán 19. 06. 2020
- There has been a LOT of buzz and over hype around data AI, data science, machine learning, and deep learning. Don't get me wrong, there are some new methods that could enhance and improve what we do in finance, banking, trading, and business in general. AI and CS (computer science) themselves have drastically changed the world we live in. I try to sit in a realistic middle ground of being optimistic of what these new methods can do but also being realistic of the current limitation.
This video will cover some of the definitions to start and sadly most people don't agree on the definitions but the ones provided are fairly well accepted. Then I cover examples of plain AI in banking (data quality checks), a discussion of machine learning and statistics, some deep learning, and don't forget ethical concerns.
Deep Learning with Python (my favorite neural network book with affiliate link):
amzn.to/3cTLLXi
Many of the charts and imagines come from Deep Learning with Python. I highly recommend this book as an introduction to neural networks and how to implement them. I would also recommend you spend time learning the math behind neural networks. Below are a few videos of me explaining a simple MLP NN.
Neural Networks (Easy Introduction):
• Neural Networks (Easy ...
Neural Network Math: Forward Propagation:
• Neural Network Math: F...
Neural Network Math: Backward Propagation:
• Neural Network Math: B...
SUPPORT THE CHANNEL:
Quant t-shirts, mugs, and hoodies:
teespring.com/stores/fancy-quant
Connect with me:
/ dimitri-bianco
/ dimitribianco
☕ Show Your SUPPORT and Buy Me a COFFEE ☕
ko-fi.com/fancyquant
Highly agree :) echoing the sentiments of Ellenberg's 'How to not be wrong' , that one needs to actually understand the assumptions of the models and be careful. Great video!
This content should be used by universities for extra activities
how is this comment here from 1 month ago when this video was uploaded today
@@sentralorigin DATA SCIENCE!
@@sentralorigin I load videos early but I publish them as unlisted before scheduling a date. Andres actually found the video early somehow and messaged me about it.
surprise!
Hi Dimitri,
I'm a physics student working on CNNs as bachelor thesis and currently applying for a risk management internship, hopping to start my career in finance.
I don't have words to express how useful and inspiring your videos are to me, and surely for many more people. Thank you!
Thanks for being a part of the channel! Feedback from subscribers like you help shape the content.
Thank you for the upload.
Thanks Dmitri!
The combination of GBMs followed by an ARIMA 'layer' has turned out to be a good combination for Time Series prediction, interesting and also a proof that you should always test the old and the new
Interesting. Thanks for the insight.
Hi Dimitri, I really enjoy your content and find it very informative.
I have recently started working as a data scientist in an insurance company and previously work for a short time in banking. I have seen things like GBMs been used to measure claims and default rates in conjunction with "traditional" statistical methods. On the other hand there seems to be a trend with a few hedge/quant funds looking for NLP specialists, which seems to go more into the "non-traditional" methodologies.
Have you thought about doing a video covering these topics? It would be interesting to know your opinion on what specific use cases data science is going to find to solve challenges in the finance. Will ML completely replace these traditional statistical methodologies? What other non-traditional use cases data science will find in finance?
I'll consider making a video around this idea. You've hit the nail on the head with the big advantage of AI/ML, it's finding new problems that weren't solvable by traditional methods in the past. The reason hedge funds are using NLP is to try and find an area where information hasn't been extracted that could add value. They are very secretive on where they are pulling data from in an attempt to beak competitors.
Bookmark: 28:00
Hey Dimitri ,
Thanks for taking all this effort making video to encapsulate the hype that has been going on data science/ML these few years , and taking the time to explain what it really is through easily understood method like using a Venn diagram . It seems like universities are coming up with new degrees courses like (data science with AI , data science analytics ) , and I'm alittle sceptical and wary . Do you feel it's going to be detrimental to students , instead of them doing a pure stats degree with a minor in comp science ? Seems like university marketing cashing in on the new hype , but I might be wrong .
The whole data science concept is a bit interesting as I view it as computer scientists doing statistics. They have created some cool new tools which has been very helpful for some unique problems like image and language processing. I also wonder what will become of data science. A lot of it is really just traditional statistics meaning the skills and methods overlap a lot with statistics jobs and education. For the banking industry I think data science will be brought under the umbrella of quantitative finance. In other industries I can see them trying to keep them separate however I think this will cause issues in the long-run. For example, who get what projects and how do you decide what team can model it better?
@@DimitriBianco Tremendous respect Dimitri for making many kinds of videos( career advice , time series modelling etc ) to give many an insight of being a quant .
I'm just a grease monkey in the oil field , but I like to watch your videos ,albeit some beyond my knowledge . I've been busy last few years going to rig and working , now I'm more free to watch videos to learn more and a general knowledge of development going on in the world .
I'm not surprise you are so good in what you do and highly successful in the quant field . As the saying goes , you share knowledge you get knowledge as well .
Hey Dimitri, what are some good books to read (preferably on the shorter end around 200 pages) about Data Science and Machine Learning in business? Right now I'm reading Machine Learning in Business by John C Hull (professor at a nearby university to me, University of Toronto) which serves as a nice introduction to the subject.
I typically read larger textbooks so I'm not sure what to recommend for shorter books. Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow by Geron is a great introductory book however it's 800+ pages.
Bookmark: 18:10
Hi Dimitri, could you suggest a good API to get almost real-time (every 15 minutes) prices and volume and same time interval for historical data?
I have not seen a good API for pricing and volume data. Quality data is very expensive. Banks and investing firms spend in the tens of millions every year for data.
In your opinion, how difficult is it for a data scientist to become a quant and vice versa?
This is a tricky question. If you have all the skills needed to do quantitative finance and you're a data scientist you should be able to switch over and vise versa. The reason this is a tricky question is because there are a lot of people who think they have the education but realistically lack the basics. In banking I foresee data science being merged into the traditional quant roles. For example, I am already working with data science models and methods as a quant. A lot of the quant masters programs have already incorporated data science as well. I'm seeing more programs spend less time on derivative pricing (stochastic calculus for example) and more time on data science.
I don't think it makes sense to have two separate teams either. If you had a problem that needed to be modeled, how would it be assigned? Realistically someone should review the problem and work with the data a bit to figure out what method would work best. By that point a good portion of the project would have already been done by one team.
Hi Dimitri, I'm currently a first year student double major in finance and statistics. I wonder could I get a quant related job after graduate from university, or I need to find some other jobs and get a master degree to be qualified?
If you want to work in quant finance in the US you'll need at minimum a masters. Other countries are less competitive and some of them don't require a graduate degree.
Nice
chao ca nha❤️