News, Views & Insights
Announcing Wolfram Finance Platform
One key project for me recently has been the new Wolfram Finance Platform, which I am pleased to announce today.
This is a major new initiative for us to create the ultimate computation environment for finance. It builds on our existing computational technology with extra capabilities and professional support services.
As part of this, I spent some time interviewing finance customers in the city of London about what they liked and didn’t like about Mathematica, what they wanted, and why some of their colleagues didn’t use it.
It turned out to be an exercise in staying open-minded and listening—because while we are naturally most excited about all the great computation that we provide (including “finance” functionality), time and again the issues that mattered were the less glamorous workflow improvements.
We have been focused on improving workflows more generally, for example supporting great data analysis, statistics, and visualization with highly automated import and export and reporting. The flow from data to analysis to CDF report is now a smooth process.
But looking at the specific needs of one industry segment reveals further optimizations. The new Bloomberg feed link is one such example. Being able to flow live, trading quality data directly into a computation or to create a CDF where the charts update with the market has now become trivial because we took that extra step to make the connection seamless.
Applying one of our our core principles—to automate as much as possible—makes this feature particularly powerful. The link includes a parameter discovery interface that automatically generates API code. This means that, armed with a new trading idea, you can prototype fast with our analysis capabilities and just paste in the generated code to call to the live data, and you can be ready to trade your new idea in minutes. Feed that into some of the built-in visualizations, and you can have a live CDF-powered dashboard to watch in just a few more minutes. One message I have been hearing repeatedly is that this kind of “algorithmic agility” is a key competitive advantage in finance.
Check out the new Wolfram Finance Platform, and watch this space for further improvements to both finance functionality and other optimized workflows that are in Wolfram Finance Platform‘s development pipeline.
Have you made the date functions any faster? This has been the greatest weakness I have found in using Mathematica in finance; the date functions are two orders of magnitude faster than any other platform I have tested. When dealing with empirical data, it can be crippling.
argh, typo above. Mathematica’s date functions are two orders of magnitude *slower* than any other platform I have tried.
Date functionality is being updated, with one of the issues being performance improvements. However that project was not ready for this release.
What’s the difference between financial platform and a normal mathematica license? Is it a bundled support option or is there something more?
There are differences in both services and technology.
The services include 24/7 support, included consulting and training time and access to developers.
On the technology side there is Bloomberg feed access, and greatly enhanced CUDA based FinancialDerivative functionality as well as some other bundled technologies such as Excel linking, and time series, with more on the way.
Michael, consider using JODA Time, the Java date and time API, http://joda-time.sourceforge.net/, and then call from M with JLink. It’s BLAZINGLY fast, and extremely flexible. They use key concepts, such as instant, partial, interval, duration, period, chronology, … very flexible modeling, and beats M by magnitudes speed-wise. If you look at the M code for the Calendar` package, no wonder this takes forever. I have “outsourced” all my date and time functions to JODA.