LineUp: Visual Analysis of MultiAttribute Rankings
Samuel Gratzl, Alexander Lex, Nils
Gehlenborg, Hanspeter Pfister, and
Marc Streit
Background
• Rankings
– Structuring unorganized collections of items.
– Computing a rank for each item based on the
value of one or more of its attributes.
• Usage of Ranking
– To arrange tasks or to evaluate the performance of
products relative to each other.
• University ranking
• Restaurant ranking
• Laptop ranking
Background
• Problem
– The visualization of a ranking itself is
straightforward, but its interpretation is not.
• How there attributes contribute to the rank.
• How changes in one or more attributes influence the
ranking.
– To interpret, modify, and compare such rankings.
• Advanced visual tools are needed to make this process
efficient.
Contributions
• A new technique that addresses the
limitations of existing methods.
– A comprehensive analysis of requirements of
multi-attribute rankings considering various
domains.
– The design and implementation of LineUp, a visual
analysis tool for creating, refining, and exploring
ranking based on complex combinations of
attributes.
Requirement analysis
• Encode rank (1/10)
– Users of the visualization should be able to quickly
grasp the ranks of the individual items.
• Encode cause of rank (2/10)
– In order to understand how the ranks are determined,
users must be able to evaluate the overall item scores
from which the ranking is derived and how they relate
to each other.
• Support multiple attributes (3/10)
– Users must be able to combine multiple attributes to
produce a single all-encompassing ranking.
Requirement analysis
• Support filtering (4/10)
– Users might want to exclude items from a ranking
for various reasons.
• Enable flexible mapping of values to scores
(5/10)
– The ranking visualization must allow users to
flexibly normalize attributes.
• Adapt scalability to the task (6/10)
– There is a tradeoff between level of detail and
scalability.
Requirement analysis
• Handle missing values (7/10)
– As real-world data is often incomplete, it must be
able to deal with missing values.
• Interactive refinement and visual feedback
(8/10)
– To enable users to judge the effect of
modifications which can influences on rankings.
Such as dynamically add and remove attributes,
modify attribute combinations, and change the
weights and mappings of attributes.
Requirement analysis
• Rank-driven attribute optimization (9/10)
– Optimizing the settings (values, weights) to find
the best possible ranking of a particular item.
• Compare multiple rankings (10/10)
– An interactive ranking visualization that fulfills 1-9
is a powerful tool addressing many different tasks.
However, in some situations users are interested
in putting multiple rankings into context with each
other.
Multi-Attribute Ranking Visualization
Technique
• LineUp
– An interactive technique designed to create, visualize, and explore
rankings of items based on a set of heterogeneous attributes.
– The visualization uses bar chars in various configurations for
ranking.
Illustration of different ranking
visualization techniques
LineUp
• Vedio
• http://www.youtube.com/watch?v=iFqCBI4T8
ks
Evaluation
• User Study
– Eight participants (6 male, 2 female) between 26
and 34 years old. They are all researchers or
students with a background in computer science,
bioinformatics, or public health.
– Compare with Excel and Tableau.
– Most of them were convinced that LineUp would
save time and allow them to gather more insights.
Questions ?
Thanks