Algorithmic trading is the process of using computers for trading. The speed and frequency that is possible through algorithmic trading cannot be matched by humans. But this kind of black box trading can make investors nervous, especially when large amounts are at stake.
This is precisely where data science comes into picture. Savi Basavaraj, head of algorithmic trading at Edelweiss said “Data science in a way sits on top of algorithmic trading, to give the necessary controls, insights, intelligence and visibility to the traders”. He was speaking at the data science congress summit 2017 held in Mumbai, India.
Edelweiss is a diversified financial services with headquarters in Mumbai, India. Savi was accompanied with Madhur Shukla, data scientist at Edelweiss. Madhur explained that data science is used in algorithmic trading in two ways: pre trade analysis and post trade analysis.
Pre trade analysis included forecasting and simulation of various algorithmic models to decide the best fit for each investment, while post trade analysis involves measuring the success, comparing results to re-define investment strategies and improving algorithmic models by applying machine learning techniques.