Loading...

Portfolio Analytics For Algorithmic Traders

/, Finance, Optimization/Portfolio Analytics For Algorithmic Traders
Portfolio Analytics For Algorithmic Traders 2017-11-16T19:36:33+00:00

Project Description

Project Brief

Imagine you are the portfolio manager (aka ‘PM’)  of a small trading group trading in hundreds of millions of USD per day. You have thousands of positions spread across multiple asset classes.

At any point in time you need to easily and quickly have a high-level view of your outstanding positions, risks, outliers, and performance. This is precisely the type of problem we love most as it combines data visualization and statistical inference with high performance ‘too-important-to-fail’ systems engineering.

We developed a server application designed to be deployed on site (for security and performance) which tracked and displayed the following key metrics:

  • Positions across risk factors according to a customizable factor model.
  • Risk metrics such as TVar.
  • Outlier reports on daily, weekly and monthly scales.
  • Trailing sharpe ratios on daily, monthly and annual time scales.
  • Real-time asset allocation across distinct strategies.
  • Multicurrency awareness and currency risk metrics.

The data visualization was done in a personalized and easy to use manner.

Skills Needed

An intimate familiarity with the quantitative investment world was required which our background in quant trading readily provided. In addition to a deep understanding of financial risk metrics and statistics, the project had a strong data visualization and application development component.

Project Management 92%
Statistics 92%
Web Technologies 96%
Data Visualization 98%

Process

Initial Concept Planning

Although the needs of a PM in the quant trading world are fairly universal, we wanted to make sure that the final system would have the functionality and look tailored to the client at hand. Mockups of the required dashboards were made first, and a final subset was selected. This is the kind of project where most of the design work can be graphically and from the get-go, before touching a single line of code.

Drafts & Revisions

Upon approval of the plan, the a prototype incorporating the agreed upon designed was implemented and deployed. User testing and feedback was collected and used to iteratively improve the system.

Final Delivery

The final system was implemented and deployed. This time with full test coverage, given the ‘too-important-too-fail’ nature of the project (mistakes in the report dashboard could potentially cause large financial losses – so extra rigour was required). After a final round of live testing the system was incorporated into the daily workflow of the PM.

Let’s Get Started