Advance Technology and Methods within Finance

Precipitated by the recent acknowledgment by an award being given to the company Fractal Edge for its innovative work in financial visualization, and my, timely introduction to a gifted “techno-artist” named W. Bradford Paley, I was compelled by the opportunistic, and appropriate, reason to address the ever challenging duty of delivering meaningful advancements in the field of financial visualization. I must admit that before I met Bradford that I was only aware that I understood pictures more than a long string of lexical and syntactically correct text describing an abstract concept. But his passion and intuitive style of presenting visualization techniques revitalized my spirit, and my belief that the “Holo-Deck” was not “crap”. That there is indeed much more available when you put your mind, or cerebral cortex, to it; which he does to our gratitude!

Many cognitive studies have demonstrated convincingly that human cognition relies more heavily on visual than numeric stimuli. People absorb and process large amounts of visual data using the visual cortex, a specialized area of the brain, every second. Despite this fact, the most common basis for presenting and analyzing quantitative research in financial engineering is still numerical, e.g., spreadsheets and tables. Some have argued that visualization is less achievable for financial applications because of the high dimensionality of the problems. These arguments have been refuted dramatically in other highly quantitative disciplines such as fluid dynamics, electrical engineering, mechanical engineering, molecular biology, and meteorology, in which visualization has come to play a central role in both basic research and industry applications. 

Over the past decade our understanding of how we precive data and process it has improved and so has the corresponding computer science field of information visualization. It has developed and grown over the last few decades, enabled by the capability of economically massive storage, powerful 64-bit processors, GPUs with capabilities sometimes more impressive than their CPU, and new methods to manipulate the data using high performance vector and Hilbert-based databases. Information visualization hinges on the idea that data can be visually represented in useful and insightful ways that enhance our understanding of the information. The flip side is that complex data can often be grasped quickly and intuitively when rendered visually. An effective approach to realizing theses innovative visualization techniques is to bring to bear the multiple disciplines of cognitive science, graphic design, computer science and, of course, Finance.  

Admittedly, there have been financial companies in the data visualization space for a sometime now. The heat-maps visualizations are perhaps the most popular on trading desks. A popular one found on smartmoney.com uses a large rectangle to represent the market capitalization of an entire industry. Within the industry rectangle contain smaller rectangles that represent sectors within that industries market capitalization. Further inside the sectors are rectangles that represent the market capitalization of individual companies. The sizes of these rectangles are dynamic and change in size relative to the price changes of the individual stocks in real-time. All the rectangles sizes are normalized so they continue to fit into the larger rectangle perfectly.  Companies that are down in market capitalization are shaded red and ones that are up are shaded green. This goes for the sectors as well. One gets an intuitive understanding of the state of the market quickly and easily. The value of this simple tool is impressive when one considers that you are only utilizing two of René Descartes Cartesian coordinates, several shades of color, and some careful thought.

Treemapping is a method for displaying information about entities with a hierarchical relationship, in a “space-constrained” environment (such as a computer monitor).

 –Wikipedia

Treemaps can be used to track the performance of portfolios or a desk’s risk exposure more quickly and seamlessly than existing alternatives. Users can track the relative performance of, say, 1,000 stocks in real time and drill down to get more detailed information, instead of looking at those 1,000 stocks in a 30-page spreadsheet. JPMorgan uses heat-mapping toolkit to produce Credit Map, a visualization product that enables the bank’s institutional customers to visualize real-time developments in the credit markets.

Now, to the award winning product: Fractal:Edge. Fractal:Edge’s visualization is vaguely similar to the heat-map’s with some notable improvements. In essence it uses the moments of fractals to represent varying market concepts. In the figure below, for example, the company’s real-time chart of the global equities market is represented by a large circle with progressively smaller ones contained inside, akin to a Venn diagram. In this depiction, the regions of the world are shown as large circles. Within those circles are rings containing what look like smaller spheres. The rings represent indexes in that region and the spheres within each of them depict the sectors within that index. Each circle, ring and sphere is color-coded based on performance, and a viewer can drill down from a global level to the individual stock level in moments.

Fractal:Edges’ Fratal Map

Less know, but more impressive is work done by W. Bradford Paley, professor at Columbia University that has both sides of his brain functioning in, undeniably, perfect unison, creating innovative visualization methods intuitively. Out of several impressive visualization techniques he has created one stands out for me; it is TextArc. First, it converts temporal data into a Euclidian space with vector features. How it works is that a time-series of data is “rolled out”, line by line around in an oval outline. Secondly, an inner arc is drawn by analyzing the unique elements in the outer arc. When a new data element is discovered in the outside arc, its datum is set where it was found. The method then continues to parse the outside arc searching for the same data element, when it finds another instant it “pulls” the original instant towards the new instance in a vector type manner, some finite distance. This process continues until the out side arc is exhausted and the method moves on to the next unique element. The original element might have been “pulled” in many different directions resulting in its ending state being almost anywhere within the oval. Paley then labels the data element with its real value and shows the array of vectors emanating form the element, like rays of light from a star, some small finite distance. The result is astounding. Paley has worked in the finance space for several years now and I’m hoping to see some real impressive applications to his work.

TextArc by W. Bradford Paley

Finally, let’s look at enabling technologies for visualization: and let’s keep our feet on the ground and shelve the “Holo-deck for today”. A few years ago I encounter a very interesting piece of hardware at an SIA show that I have not seen since. It was a monitor akin to a simple double-glazing window found in your home. The first panel was, seemly, a normal LCD panel, the second panel, which was laid, directly over the first was a translucent panel. That was it, the rest we leave to Newton and his description of the properties of “Opticks”. You see the magic of this device simply lies in the properties of geometry and stereoscopic vision that creates a profound visual effect that is actually quite normal and performed by our eye ever few seconds.

One way to describe this physical behavior is the use of a term called  Depth of Field and is described as the following:

In optics, particularly film and photography, the depth of field (DOF) is the distance in front of and behind the subject which appears to be in focus. For any given lens setting, there is only one distance at which a subject is precisely in focus, but focus falls off gradually on either side of that distance, so there is a region in which the blurring is tolerable. This region is greater behind the point of focus than it is in front, as the angle of the light rays change more rapidly; they approach being parallel with increasing distance.

–Wikipedia

You can imagine that this new two paneled monitor uses DOF to its advantage. With its distance from your face at only 12” to 24”, and with a normally symmetrical human operator’s eyes being 2” to 3” apart, creates a 12 degree horizontal field of view. Ignoring the math for now this yields a very small and precise DOF of around ¼ of an inch. This DOF is more than enough to be able to focus on the back or front panel only, and eliminating the unfocused panel within your field of view. The demonstration I saw had both panels with Excel sheets filled with a massive amount of numbers; it was very easy to read one panel or the other. Imaging the power of this technology coupled with some real innovative visualization techniques; one can only think.

Another why to describe this physical behavior of the dual-panel monitor is the use of the more formal phyiscs field of stereopsis. Depth from stereopsis arises from the slightly different positions each eye occupies on the head, a form of parallax, as described above. A stereoscopeis a device by which each eye can be presented with different images, allowing stereopsis to be stimulated with two pictures, one for each eye. This has lead to various crazesfor stereopsis, usually prompted by new sorts of stereoscopes. Stereopsis appears to be processed in the visual cortex in binocular cells having receptive fields in different horizontal positions in the two eyes. Such a cell is active only when its preferred stimulus is in the correct position in the left eye and in the correct position in the right eye, making it a disparity detector. When a person stares at an object, the two eyes converge so that the object appears at the center of the retina in both eyes. Other objects around the main object appear shifted in relation to the main object. In the following example, whereas the main object remains in the center of the two images in the two eyes, the cube is shifted to the right in the left eye’s image and is shifted to the left when in the right eye’s image. Because each eye is in a different horizontal position, each has a slightly different perspective on a scene yielding different retinal images. Normally two images are not observed, but rather a single view of the scene, a phenomenon known as singleness of vision. If the images are very different (such as by going cross-eyed, or by presenting different images in a stereoscope) then one image at a time may be seen, a phenomenon known as binocular rivalry.

Regardless of which description helps you understand this application the simple fact remains that the monitor is very innovative and impressive.

With profound advances in enabling technologies and a “renaissance” in thought process, data visualization is starting to gain traction in the financial services industry. As this happens, the adoption of visualization tools is likely to be rapid, particularly as automated and electronic trading takes off and market data, like Level II depth, become more accessible.