Article first published as Many Eyes: Data Visualization with IBM and Cognos on Technorati.

Probably the little-known champion of easy data visualization is Many Eyes, a SaaS offered via Cognos and IBM. Hosted on IBM’s servers, Many Eyes allows pure data in either text or table format to be uploaded and processed into multiple visualization formats.

The Many Eyes IBM/Cognos system garners user input in either free-text or spreadsheet format, stores it on the server, and makes the data-set open to any user of Many Eyes. The tasks here are multiple: one major one is allowing free information access, another is visualization creation, a third is elicitation of user evaluation via comments and feedback, and yet another is gathering sources of free data for IBM’s corporate use. Essentially, using Many Eyes is an exchange of data for access to the visualization service.

Many Eyes is very clear in the visualization creation process, making each step clear: first data collection takes place, then data abstraction, then visual abstraction, and then visual delivery of the transmogrified data. Importantly, it is a static system; once the data is uploaded, it cannot be modified, only visualized in multiple ways.

IBM makes several solid assumptions about users, many of which are typical of visualization developments: they prefer a neutral or pastel color schema, they are inputting either free text or spreadsheet data, and they are making no mistakes with their data source. The latter is particular to information visualization, which is essentially a complex mathematical model.

IBM effectively has built an iterative system where users give feedback to other users, thus giving IBM a high-quality visualization analysis system. While IBM could conduct numerous investigations and evaluation studies in an empirically validated setting, they have chosen to crowd-source both their data sources and visualizations. This stands out as a particularly effective research model, acting as a type of rough capture system and aiding in not only metric analysis, but qualitative meta-data collection. As well, IBM is experimenting with pseudo-open source methods and community data collection.

Many Eyes will remain a pseudo-open source system until it is available in multiple places without the looming threat of restricted access. IBM ultimately retains control of the visualization system and data, unlike a user running Prefuse Flare, where the data can be manipulated on the client’s computer and disseminated without a corporate intermediary. An alternative to Many Eyes would be the software Tableau, or building an ontology via the Protégé system and then processing it in the Protégé module TGVizTab.

In total, Many Eyes is able to render more complex data, but requires a higher level of data sophistication and purity of data input. When you need to create stunning interactive pictures quickly, IBM’s Many Eyes is a great solution. After all, in the 21st century, knowledge is art.

Many thanks to Prof. Chaomei Chen at Drexel University‘s iSchool for his depth of visualization knowledge. Image courtesy of Closeup Eye by Petr Kratochvil at Public Domain Pictures.