Finance is a historical sector in which data is present. We will see how it has always allowed this sector to make decisions and the evolutions that have taken place.

‍

Data collected since the beginning of time

‍

Let's start with the example of the stock market, with the data with the highest volume: stock prices. A company like Euronext, which is the main stock exchange in the euro zone, will record all the transactions that have taken place in the form : 

• Buyer
• Salesman
• Price 
• Volume (number of titles)
• Time of the transaction (to the nearest thousandth of a second or even to the millionth)

Millions of such transactions take place every second. We must be able to process them in real time, in the right order and with the right rules in place, respecting in particular the balance between supply and demand. So historically, this is a case where we have data very naturally, but where we also arrive very quickly at Big Data, since the volumes of data collected are very large and we must be able to process them quickly.

In addition, all the historical data of companies form significant volumes of data with sales, profits, margin, assets. And also textual data such as perspectives, strategy, objectives... This data will allow to value a company at a given date and to forecast its potential evolutions.

‍

The evolutions brought by the advances of the data

‍

This list is obviously not exhaustive, but here are some of the topics on which data and machine learning have allowed the world of finance to evolve.

Real time

Maybe you didn't live in those days, but most individuals followed the stock market on a daily basis by buying their newspaper or making a phone call - often for a fee. Bringing in real time helped the financial world to grow. What helped there was the ability to process data quickly.

Another piece of information linked to real time comes from machine learning. We can analyze social networks to understand the political and social trends, the feeling of the populations towards such or such event or company. We can, for example, find out which company is the most talked about on social networks. On the other hand, when we are interested in a specific company, we can look for what people think about it.

Fraud detection

Fraud detection is a subject for which new techniques in data have made it possible to advance. It allows to detect suspicious behaviors in transactions. A person has a certain purchasing habit and suddenly starts spending large amounts of money in an unusual way? Fraud is then possible and we can warn the right people to avoid financial losses.

For banks, this subject is important for what is called KYC (know your customer) - the bank must show that it knows its customers well to financial institutions - but also for its reputation with its customers.

Improvement of risk models

Using data to make certain financial decisions is a good thing. It avoids emotional decisions. It reduces the risk of bad investments and bad payers. We can think for example of the company October which has automated part of its risk models to propose credits. Where a traditional bank can take several months to offer a loan, this company can do so in barely a week for loans ranging from a few tens of thousands of euros to a few million euros.