Data visualisation software rises from the ranks

Researchers from Harvard University have created a piece of open-source software which enables users to analyse and rearrange ranking systems based on the attributes they find the most important. Their paper on the software, published in IEEE Transactions on Visualization and Computer Graphics, won the best paper award at the InfoVis conference in October 2013.

“Rankings are a popular and universal approach to structuring otherwise unorganised collections of items by computing a rank for each item based on the value of one or more of its attributes,” the researchers said in the paper.

“While the visualisation of a ranking itself is straightforward, its interpretation is not, because the rank of an item represents only a summary of a potentially complicated relationship between its attributes and those of the other items,” they continued. And this does not even take personal bias and subjectivity into consideration.

The researchers thus proposed LineUp - “an interactive technique to create, visualise and explore multi-attribute rankings”, according to Alexander Lex, a postdoctoral researcher at the Harvard School of Engineering and Applied Sciences (SEAS). Lex created the software along with his adviser Hanspeter Pfister; An Wang, Professor of Computer Science at SEAS; Nils Gehlenborg, a research associate at Harvard Medical School; and Marc Streit and Samuel Gratzl at Johannes Kepler University.

“This interactive technique supports the ranking of items based on multiple heterogeneous attributes with different scales and semantics,” the researchers said. Configured as a table-based bar chart, with columns for each attribute, “it enables users to interactively combine attributes and flexibly refine parameters to explore the effect of changes in the attribute combination”.

Lex described how prospective students might use the system to analyse the rankings of universities. They may filter rankings to their own country and then put more weight on the strength of the subjects they are interested in studying, as well as the faculty-student ratio. They can even invert items with lower rankings - eg, if they aren’t interested in studying arts, they can give higher rankings for universities with lower values for arts subjects.

“Columns can be moved, resized, removed and restored,” Lex said. When ranks change, animations and colour indicate those changes. To better keep track of big changes and alternative weighting systems, the table can be split in two, allowing for side-by-side comparisons.

LineUp was actually created as part of a larger software package called Caleydo, a framework which visualises genetic data and biological pathways, eg, to analyse and characterise cancer subtypes.

“LineUp really was developed to address our need to understand the ranking of genes by mutation frequency and other clinical parameters in a group of patients,” said Pfister. “It is an ideal tool to create and visualise complex combined scores of bioinformatics algorithms.”

But the software can be used for all sorts of rankings - even predictive ones, by applying ‘what if?’ scenarios. It thus equips people without deep statistical or technical knowledge with a data analysis tool for everyday life.

LineUp is available as open-source software at http://lineup.caleydo.org. The team is currently working on a web-based version.

Source