Our journey in this book

This book is not only about investing or trading in the finance sector; it's much more as a direct result of the love story between computers and finance. Investment firms have customers, often insurance firms or pension funds, and these firms are financial services companies themselves and, in turn, also have customers, everyday people that have a pension or are insured.

Most bank customers are everyday people as well, and increasingly, the main way people are interacting with their bank, insurer, or pension is through an app on their mobile phone.

In the decades before today, retail banks relied on the fact that people would have to come into the branch, face-to-face, in order to withdraw cash or to make a transaction. While they were in the branch, their advisor could also sell them another product, such as a mortgage or insurance. Today's customers still want to buy mortgages and insurance, but they no longer have to do it in person at the branch. In today's world, banks tend to advise their clients online, whether it's through the app or their website.

This online aspect only works if the bank can understand its customers' needs from their data and provide tailor-made experiences online. Equally, from the customers, perspective, they now expect to be able to submit insurance claims from their phone and to get an instant response. In today's world, insurers need to be able to automatically assess claims and make decisions in order to fulfill their customers' demands.

This book is not about how to write trading algorithms in order to make a quick buck. It is about leveraging the art and craft of building machine learning-driven systems that are useful in the financial industry.

Building anything of value requires a lot of time and effort. Right now, the market for building valuable things, to make an analogy to economics, is highly inefficient. Applications of machine learning will transform the industry over the next few decades, and this book will provide you with a toolbox that allows you to be part of the change.

Many of the examples in this book use data outside the realm of "financial data." Stock market data is used at no time in this book, and this decision was made for three specific reasons.

Firstly, the examples that are shown demonstrate techniques that can usually easily be applied to other datasets. Therefore, datasets were chosen that demonstrate some common challenges that professionals, like yourselves, will face while also remaining computationally tractable.

Secondly, financial data is fundamentally time dependent. To make this book useful over a longer span of time, and to ensure that as machine learning becomes more prominent, this book remains a vital part of your toolkit, we have used some non-financial data so that the data discussed here will still be relevant.

Finally, using alternative and non-classical data aims to inspire you to think about what other data you could use in your processes. Could you use drone footage of plants to augment your grain price models? Could you use web browsing behavior to offer different financial products? Thinking outside of the box is a necessary skill to have if you want to make use of the data that is around you.