Visualize Your Data
Data visualization is not a new concept. Anyone who ever prepared a business presentation knows that a colorful diagram expressing the data visually has a much greater impact on the audience.
But data visualization is more than merely a presentation technique. With data visualization you can easily gain new insights, which enable you to take action. Coming up with the same insights using only the raw data is sometimes much more challenging.
Take for example the fascinating service provided by Eyetools. Using eye tracking technology and a focus group you receive a heat map of your web site. The heat map visually presents the things users saw on your site. What they literally saw.
The red areas in this heat map, for example, indicate that almost all subjects halted their gaze for at least a fraction of a second at this part of the page. “Blank” areas, on the other hand, were never seen by the subjects.
The interesting thing about this service (apart from the technological aspect of course) is that with a single visual image you get immense amount of information, which you can use to take immediate action. Trying to capture all this information in any other form would have resulted in a flood of information, which you would have found difficult to manage.
Obviously, this innovative visualization is much more effective in this context than a simple graph based on raw data. The beauty of the heat map is that it relates to the actual subject of the research. The visual data gathered during the research overlies the actual web page it refers to. This scheme makes the data self explanatory.
The same idea can be applied in different contexts in the world of software development management.
Yesterday, for example, I saw a demo of a customized defect tracking schema built on top of a well-known issue tracking tool. Part of the schema enabled a detailed breakdown of the product to a hierarchy of components and sub-components. Now, I am not suggesting that such a large amount of data is necessary for the purpose of defect tracking in most projects. But in light of the complexity of these projects, this schema seemed appropriate.
The problem was that a project manager trying to get a sense of the status of defects in his project might find himself overwhelmed with information. A simple project-level query might result in hundreds of records, which might obscure the real picture the project manager is trying to get.
This is where data visualization can make the difference. Since the structure of the product (in terms of components) is already part of the metadata in the system, generating a visual representation of a query is relatively easy. Forget the simple bar charts. Why not generate a block diagram of the product, where each component and module is colored according to the number of open defects.
With such an image, the project manager can instantly see where action is most needed. He can easily identify which critical components are in bad shape, or which components are suffering from recurrent functional problems. The same image could be spiced with additional layers to reflect the severity of defects, history of defects, etc.
Understanding the real status of a complex project from such a visual is by far easier than using the raw data for this purpose.
Of course this is just one example. I bet you can find other ideas for innovative data visualization in the same spirit.
