The origins of stock exchange, how we know it nowadays, go way back to the French Revolution of the late XVIII century. Nowadays, stock markets, where all kinds of financial products are negotiated (Gold, Brent Crude, Rice, Debt, Forex, etc…), generate daily 2.5 quintillion bytes of data, which represents a unique opportunity to process, analyze and leverage information in a useful way.
Technology is escalating at an exponential rate and the consequential advantages are far-reaching. As a consequence, the big quantity of data handled everyday is transforming the way of operating in industries and in the financial sector.
Machine Learning and algorithmic calculation are used more and more in financial markets in order to process all information generated daily and to be able to make predictions and take decisions we humans are not able to.
When it comes to Big Data, which type of information should we analyze?
Social media with sentiment analysis, real-time news, history of transactions or financial products, etc. To make a real example, recently we viewed how Twitter can impact stock prices on actions like Tesla’s.
It is estimated that ¾ of moves produced by markets are conducted by algorithms.
On May 6th 2010, we experienced the known “Flash Crash”, in which the American Dow Jones suffered a drop by 9%, which was recovered in just 2.45 minutes. In April 2015, Navider Singh Sarao, a London-based operator, stalled. Apparently, they used a program used to generate big selling orders to force prices down. Consequently, they cancelled selling orders and bought it at the lowest fare on the markets.
Source: Bloomberg/The New York Times
During this lapse, one can see the quantity of operations produced by many investment funds using their investment robots to seize market opportunities.
While managers were deciding what to do, computers were calculating risks and actualized purchasing and selling in fractions of seconds.
Therefore, the analysis of big quantities of data is influencing financial markets and on industries in general. These are the most relevant points:
Big Data analysis on financial models
Financial analysis is not only about prices and behavior analysis. Now we can analyze anything affecting prices, like social and political trends, news etc.
Big Data analysis can be used with predictive models to estimate the rate of return of our investment. The access to this quantity of information paves the way to predictions with a higher degree of precision in order to effectively manage risks.
Technology today allows us to receive all this information in real time and to be able to take similar decisions to the ones taken by humans at a much higher speed. This allows our decisions to be taken automatically in a much shorter time reducing manual errors arising following influences of the moment.
With tools like Microsoft Azure ML, we can analyze in real time structured data as well as not structured data, social media, share market values, news analysis, etc.
This situational analysis is very valuable, now that the share market is a easily influenced archetype.
The strength of this technology remains still unexploited and perspectives are unimaginable. Automatic learning allows computers to really learn and take decisions based on new information received learning by mistakes from the past, using algorithms.
Nowadays more and more enterprises replace portfolio management, raw materials, risks, investments, etc. with Machine Learning models. To continue, I will give some examples where a machine Learning model can be particularly beneficial:
- Stock of products or raw materials with an influence by financial markets (Raw, Cellulose, Energy, etc)
- Exposition of business by value change
- Risk management
- News impact analysis (natural catastrophes, financial news, politics)
- Sentiment analysis on social media