Research

NEW BOOK: "Derivatives Analytics with Python"

Derivatives Lecture at Saarland University

Fast Monte Carlo Valuation of American Options with Stochastic Volatility

RECENT CLIENT PROJECT

Visixion GmbH has developed and conducted a focused Python training for selected people at Eurex, one of the world's leading derivatives exchanges. The major goal is to replace in certain areas a heterogeneous IT infrastructure (including, amongst others, Matlab and R) by Python as the main programming environment. Requirements are increased productivity, fast development cycles, easy collaboration, easy-to-maintain solutions and high performance.

DEXISION In The Press

DEXISION mentioned in the March 2010 Cover Story of Wilmott Magazine as first Derivatives Analytics suite based on the Python programming language. Read the whole Story.

TECHNOLOGY

We at Visixion are a firm believer in Open Source in general and in Python (www.python.org) in particular. It seems that simple, Web-based applications are best built on this ground and that Python is perfectly suited for derivatives analytics.

Why Python for a Derivatives Analytics Suite?

Some might wonder whether Python is the right language for financial engineering and derivatives analytics since these disciplines require robust code with high performance. From our point of view, we ourselves proved the applicability of Python in these areas with the release of DEXISION.

We believe that we—as a small team who started out from scratch—wouldn’t have been able to develop and release DEXISION at all without the simplicity, compactness, performance, interoperability and maintainability of Python code. For example, an implementation in C++ would most probably have caused as much as ten or fifteen times the number of lines of code as we now have. Testing and debugging this C++ code would also have caused both more absolute and relative overhead.

Due to the powerful array manipulation library NumPy (www.numpy.org), which is implemented in C, performance of our single or double precision array operations is comparable to implementations in C++.

So we get the same performance from Python with a tenth of the code of a C++ implementation which is in addition easier to maintain and to expand. This seems to be a real benefit. And that Python is appropriate for demanding production environments show the likes of Google, Inc.

For more on Python and Financial Engineering read the Wilmott Cover Story from March 2010.

Technological Components of DEXISION

In the following, you find the major components of the Web architecture for DEXISION (as of November 2012).

  • Suse Linux 12.1 as 64 Bit operating system
  • Apache 2 as Web server
  • MySQL 5.5.16 as relational database
  • Python 2.7.2 as core language (integrated via mod_python in Apache)
  • NumPy 1.6.1 as fast linear algebra library
  • Dojo 1.6 as JavaScript framework for the GUI
  • SVG for all custom graphics
  • Sphinx for Web documentation

If you have any questions relating to the technology of DEXISION or if you would like to implement a similar architecture for yourself, just send us a message via the Contact form.