Visualizing Data with Bounded Uncertainty
Chris Olston and Jock Mackinlay
Abstract
Visualization is a powerful way to facilitate data analysis, but it is
crucial that visualization systems explicitly convey the presence,
nature, and degree of uncertainty to users. Otherwise, there is a
danger that data will be falsely interpreted, potentially leading to
inaccurate conclusions. A common method for denoting uncertainty is
to use error bars or similar techniques designed to convey the degree
of statistical uncertainty. While uncertainty can often be modeled
statistically, a second form of uncertainty, bounded uncertainty, can
also arise that has very different properties than statistical
uncertainty. Error bars should not be used for bounded uncertainty
because they do not convey the correct properties, so a different
technique should be used instead.
In this paper we describe a technique for conveying bounded
uncertainty in visualizations and show how it can be applied
systematically to common displays of abstract charts and graphs.
Interestingly, it is not always possible to show the exact degree of
uncertainty, and in some cases it can only be displayed approximately.
We specify an algorithm that approximates the degree of uncertainty to
make it displayable while minimizing the overall loss in accuracy. In
addition, we consider new data delivery paradigms that
offer mechanisms for interactive control over uncertainty levels, but
whose use may result in hidden side effects. We propose interfaces
that offer control of uncertainty levels to the user in ways that
encourage careful use of these facilities.
Conference Paper (InfoVis 2002): [PS], [PDF]. Citation: [BibTeX]
Extended Version: [PS], [PDF]