Название: Adventures in Financial Data Science: The Empirical Properties of Financial and Economic Data, 2nd Edition
Автор: Graham L Giller
Издательство: World Scientific Publishing
Год: 2023
Страниц: 512
Язык: английский
Формат: pdf (true)
Размер: 74.7 MB

This book provides insights into the true nature of financial and economic data, and is a practical guide on how to analyze a variety of data sources. The focus of the book is on finance and economics, but it also illustrates the use of quantitative analysis and Data Science in many different areas. Lastly, the book includes practical information on how to store and process data and provides a framework for data driven reasoning about the world.

The book begins with entertaining tales from Graham Giller's career in finance, starting with speculating in UK government bonds at the Oxford Post Office, accidentally creating a global instant messaging system that went 'viral' before anybody knew what that meant, on being the person who forgot to hit 'enter' to run a hundred-million dollar statistical arbitrage system, what he decoded from his brief time spent with Jim Simons, and giving Michael Bloomberg a tutorial on Granger Causality.

The majority of the content is a narrative of analytic work done on financial, economics, and alternative data, structured around both Dr Giller's professional career and some of the things that just interested him. The goal is to stimulate interest in predictive methods, to give accurate characterizations of the true properties of financial, economic and alternative data, and to share what Richard Feynman described as 'The Pleasure of Finding Things Out.'

I am an extensive user of R, Python, but also some commercial statistical software designed specifically for time-series analysis. I use Mathematica quite a bit. I strongly prefer “script oriented” analysis over “notebook oriented” analysis due to the inherent repeatability of the scripting workflow. I do some file munging in shell scripts, principally bash and zsh, with some work in Perl. I used to work in Fortran, C, C ++, and Java, but haven’t really done so for a while. Python has become my procedural language of choice. I do still use Excel on an almost daily basis. It is a powerful tool and using it doesn’t make you a bad person.

The overall theme here is there are many tools, each of which may be good for a particular task, so you need to use a lot and learn how to string their results together. In my experience, those who evangelize a single “power tool” for all tasks are likely more interested in software engineering than learning the truths hidden by Nature within data, and data scientists make terrible software engineers.

Contents:


Readership: For quantitative and data scientists, and academics in finance, as well as people who are interested in entering these professions or just generally interested in these subjects.