Bringing Form to Dynamic Financial Data:

Russell Investment Center Data Walls

We were tasked with designing and developing a suite of data visualizations to be rendered on three forty foot video walls located in the core of the Russell Financial Center in downtown Seattle.

The screens are distributed across three different floors of the building. The audience for this project is finance professionals and the video walls would by experienced by these traders and analysts as they exited the elevators or passed by one of these three hallways on route to the various meeting rooms and spaces throughout the building.

We developed four distinct visualizations that represent specific views or perspecitives on real-time financial data. The data represented here is updated every twenty seconds.

The first view visualizes the entire Russell Index as a field of grass. Each shaft of grass represents an individual stock.

Generally speaking, if the stock leans to the right it is experiencing growth during the current day and if it is leaning to the left the value for that stock is decreasing. The entire visualization can be experienced from a distance, or in detail. From a distance, a passer by can quicky gather the overall trend in the market as they would the direction of the wind pasing over a field of grass. Up close every stock is informed by real-time data and the stocks that have stand out performance during the current day are labeled. This eliminates the visual clutter that would result from labeling all of the stocks and in addition highlights certain stories within that are more relevant to the current day. There was a great deal of time put into the planning and wireframing of this project. It became very important to develop a solid understanding of the financial data and work closely with finance experts from Russell to develop a solid set of rules that will ultimately inform every aspect of the presentation.

Moving into visual design required a great deal of overlap between the design and development processes.

Design comps were developed in ActionScript/Flash in a process more similar to a prototyping exercise than a conventional visual design process.

This was important because there is no definitive visual result because the content is a moving target and we were relying on the computer to render the visuals and are trying to arrive at a result that is always changing. It was possible to screen capture the prototypes and put together something similar to a vsiual design package, but at the end of the day the prototypes themselves were much more useful as a client facing deliverable. Developing the rules is more important than comping up a single static style frame because the results need to be both meaningful and compelling under very diverse circumstances.

Ironically enough, the first step here is to generate test data that is as close to what the final feed will be as possible. This might not sound important but the data feed was not available prior to design. This is a chicken or egg situation and the visual design cannot wait until develpment reaches the point where the feed subscription is available. The feed is expensive and the client cannot determine whether or not to license the feed until they have seen a compelling visual. Sample data has to be close to the final data so that there can be something as close to a plug and play situation as possible.

The second view animated various sectors of the Russell index across the hallway. The shift in position on the X-Axis represents movement in value for the current day as well as historical data for a 52 week period.

It was critical to define the rate that these animated sectors are traveling, what density that they will be layed out on the screen and how often they appear. It is really important to start rendering these design elements into the space and emulating the environment to begin to get comfortable with the scale and starting thinking outside of the screen. It is too easy to get hung up on the way things are rendering to a single computer or in a single abstract context and losing perpective on how these things will play out in the environment. Getting these lements projected on the wall as much as possible during this phase is essential.

We also needed to determine early on how the individual sectors would be labeled and get really granular about the data sources and exacly how they will map on to the visualization. During this phase we were able to determine how literal the interpretation of the data would be. It was never the intention for this piece to be the most literal interpretation of the data and no one wanted to see a graph. It had to be a more of an artistic abstraction. That said, everyone involved was very interested in understanding exactly how the data was to be represented and that there was a strong correlation between trends in the world of finance and what was playing out on the screen.

We developed a distinct visualization for comodities that represented each commodity as a school of fish.

The movement of the school was informed by the movement of the commodity during the course of the day. The entire length of the screen represents value to time for a single day. The schooling fish visual follow the movement of the graph without ever drawing a literal graph. We were also introducing economic indicators and financial reports here, i.e. Consumer Price Index. These are represented by a trail of bubbles rising from the bottom of the screens. This made it possible to observe how the value of the commodities was responding to financial news and couch the entire experience in a loose time framework.

Finally we created a visual reprentation for global currencies. The currencies were represented by abstract flower forms and layed out on a loose map framework. The map doubled as a world clock that represented the movement of the sunlight threshold over the planet.

The scale of the individual currencies demonstrated percent change in value during the current day. Time series data for the current day was represented by a clock-like work spiral pattern that works it's way around the currency elemnet. This time series data shows the movement of the currency value over the course of the last twelve hours. It was very interesting to develop this map and determine the logic that would define the position of the sunlight threshold relative to the map. We developed a configuration file that defined the XY coordinates for each currency and did nut use the sunlight threshold as a literal representation of daylight but instead focused on open and close for each location. For example, during the summer solstice currencies that were placed far North of the equator would have very little sunlight but the open and close would remain somewhat consistent and we chose to represent the latter.

The Russell Index visualization during calibration on site, at install.
The fundamental behavior of individual stocks are captured in this diagram.
Early sketches describing the fundamental behavior and anticipating the spacial context for the piece.
This diagram demonstrates the time based behavior and transition of the overall presentation.
A style frame/comp captured during the prototyping process.
The sectors visualization captured on site.
Early sketches describing the fundamental behavior of the sectors visualization, anticipating the spacial context for the piece.
The fundamental behavior and organizing principle for individual sectors are captured in this diagram.
A style frame/comp for the commodities aspect of the presentation captured during the prototyping process.
Early sketches describing the fundamental behavior of the commodities visualization.
A style frame/comp for the currencies aspect of the presentation captured during the prototyping process.
Early sketches describing the fundamental behavior of the currencies visualization.