Aritra Dasgupta

Pargnostics: Screen-Space Metrics for Parallel Coordinates

Interactive visualization requires the translation of data into a screen space of limited resolution. While currently ignored by most visualization models, this translation entails a loss of information and the introduction of a number of artifacts that can be useful, (e.g., aggregation, structures) or distracting (e.g., over-plotting, clutter) for the analysis. This phenomenon is observed in parallel coordinates, where overlapping lines between adjacent axes form distinct patterns, representing the relation between variables they connect. However, even for a small number of dimensions, the challenge is to effectively convey the relationships for all combinations of dimensions. The size of the dataset and a large number of dimensions only add to the complexity of this problem. To address these issues, we propose Pargnostics, parallel coordinates diagnostics, a model based on screen-space metrics that quantify the different visual structures. Pargnostics metrics are calculated for pairs of axes and take into account the resolution of the display as well as potential axis inversions. Metrics include the number of line crossings, crossing angles, convergence, overplotting, etc. To construct a visualization view, the user can pick from a ranked display showing pairs of coordinate axes and the structures between them, or examine all possible combinations of axes at once in a matrix display. Picking the best axes layout is an NP-complete problem in general, but we provide a way of automatically optimizing the display according to the users preferences based on our metrics and model.

Adaptive Privacy-Preserving Visualization Using Parallel Coordinates

Current information visualization techniques assume unrestricted access to data. However, privacy protection is a key issue for a lot of real-world data analyses. Corporate data, medical records, etc. are rich in analytical value but cannot be shared without first going through a transformation step where explicit identifiers are removed and the data is sanitized. Researchers in the field of data mining have proposed different techniques over the years for privacy-preserving data publishing and subsequent mining techniques on such sanitized data. A well-known drawback in these methods is that for even a small guarantee of privacy, the utility of the datasets is greatly reduced. In this paper, we propose an adaptive technique for privacy preser vation in parallel coordinates. Based on knowledge about the sensitivity of the data, we compute a clustered representation on the fly, which allows the user to explore the data without breaching privacy. Through the use of screen-space privacy metrics, the technique adapts to the users screen parameters and interaction. We demonstrate our method in a case study and discuss potential attack scenarios.

Conceptualizing Visual Uncertainty in Parallel Coordinates

Uncertainty is an intrinsic part of any visual representation in visualization, no matter how precise the input data. Existing research on uncertainty in visualization mainly focuses on depicting data‐space uncertainty in a visual form. Uncertainty is thus often seen as a problem to deal with, in the data, and something to be avoided if possible. In this paper, we highlight the need for analyzing visual uncertainty in order to design more effective visual representations. We study various forms of uncertainty in the visual representation of parallel coordinates and propose a taxonomy for categorizing them. By building a taxonomy, we aim to identify different sources of uncertainty in the screen space and relate them to different effects of uncertainty upon the user. We examine the literature on parallel coordinates and apply our taxonomy to categorize various techniques for reducing uncertainty. In addition, we consider uncertainty from a different perspective by identifying cases where increasing certain forms of uncertainty may even be useful, with respect to task, data type and analysis scenario. This work suggests that uncertainty is a feature that can be both useful and problematic in visualization, and it is beneficial to augment an information visualization pipeline with a facility for visual uncertainty analysis.

SimilarityExplorer: A Visual Inter-Comparison Tool for Multifaceted Climate Data

Inter‐comparison and similarity analysis to gauge consensus among multiple simulation models is a critical visualization problem for understanding climate change patterns. Climate models, specifically, Terrestrial Biosphere Models (TBM) represent time and space variable ecosystem processes, like, simulations of photosynthesis and respiration, using algorithms and driving variables such as climate and land use. While it is widely accepted that interactive visualization can enable scientists to better explore model similarity from different perspectives and different granularity of space and time, currently there is a lack of such visualization tools.

Bridging Theory with Practice: An Exploratory Study of Visualization Use and Design for Climate Model Comparison

Evaluation methodologies in visualization have mostly focused on how well the tools and techniques cater to the analytical needs of the user. While this is important in determining the effectiveness of the tools and advancing the state-of-the-art in visualization research, a key area that has mostly been overlooked is how well established visualization theories and principles are instantiated in practice. This is especially relevant when domain experts, and not visualization researchers, design visualizations for analysis of their data or for broader dissemination of scientific knowledge. There is very little research on exploring the synergistic capabilities of cross-domain collaboration between domain experts and visualization researchers. To fill this gap, in this paper we describe the results of an exploratory study of climate data visualizations conducted in tight collaboration with a pool of climate scientists. The study analyzes a large set of static climate data visualizations for identifying their shortcomings in terms of visualization design. The outcome of the study is a classification scheme that categorizes the design problems in the form of a descriptive taxonomy. The taxonomy is a first attempt for systematically categorizing the types, causes, and consequences of design problems in visualizations created by domain experts. We demonstrate the use of the taxonomy for a number of purposes, such as, improving the existing climate data visualizations, reflecting on the impact of the problems for enabling domain experts in designing better visualizations, and also learning about the gaps and opportunities for future visualization research. We demonstrate the applicability of our taxonomy through a number of examples and discuss the lessons learnt and implications of our findings.

Measuring Privacy and Utility in Privacy-Preserving Visualization

In previous work, we proposed a technique for preserving the privacy of quasi‐identifiers in sensitive data when visualized using parallel coordinates. This paper builds on that work by introducing a number of metrics that can be used to assess both the level of privacy and the amount of utility that can be gained from the resulting visualizations. We also generalize our approach beyond parallel coordinates to scatter plots and other visualization techniques. Privacy preservation generally entails a trade‐off between privacy and utility: the more the data are protected, the less useful the visualization. Using a visually‐oriented approach, we can provide a higher amount of utility than directly applying data anonymization techniques used in data mining. To demonstrate this, we use the visual uncertainty framework for systematically defining metrics based on cluster artifacts and information theoretic principles. In a case study, we demonstrate the effectiveness of our technique as compared to standard data‐based clustering in the context of privacy‐preserving visualization.

Visual reconciliation of alternative similarity spaces in climate modeling

Visual data analysis often requires grouping of data objects based on their similarity. In many application domains researchers use algorithms and techniques like clustering and multidimensional scaling to extract groupings from data. While extracting these groups using a single similarity criteria is relatively straightforward, comparing alternative criteria poses additional challenges. In this paper we define visual reconciliation as the problem of reconciling multiple alternative similarity spaces through visualization and interaction. We derive this problem from our work on model comparison in climate science where climate modelers are faced with the challenge of making sense of alternative ways to describe their models: one through the output they generate, another through the large set of properties that describe them. Ideally, they want to understand whether groups of models with similar spatio-temporal behaviors share similar sets of criteria or, conversely, whether similar criteria lead to similar behaviors. We propose a visual analytics solution based on linked views, that addresses this problem by allowing the user to dynamically create, modify and observe the interaction among groupings, thereby making the potential explanations apparent. We present case studies that demonstrate the usefulness of our technique in the area of climate science.

Privacy-Preserving Data Visualization using Parallel Coordinates

The proliferation of data in the past decade has created demand for innovative tools in different areas of exploratory data analysis, like data mining and information visualization. However, the problem with real-world datasets is that many of their attributes can identify individuals, or the data are proprietary and valuable. The field of data mining has developed a variety of ways for dealing with such data, and has established an entire subfield for privacy-preserving data mining. Visualization, on the other hand, has seen little, if any, work on handling sensitive data. With the growing applicability of data visualization in real-world scenarios, the handling of sensitive data has become a non-trivial issue we need to address in developing visualization tools. With this goal in mind, in this paper, we analyze the issue of privacy from a visualization perspective and propose a privacy-preserving visualization technique based on clustering in parallel coordinates. We also outline the key differences in approach from the privacy-preserving data mining field and compare the advantages and drawbacks of our approach.

The importance of tracing data through the visualization pipeline

Visualization research focuses either on the transformation steps necessary to create a visualization from data, or on the perception of structures after they have been shown on the screen. We argue that an end-to-end approach is necessary that tracks the data all the way through the required steps, and provides ways of measuring the impact of any of the transformations. By feeding that information back into the pipeline, visualization systems will be able to adapt the display to the data being shown, the parameters of the output device, and even the user.

VIMTEX: A Visualization Interface for Multivariate, Time-Varying, Geological Data Exploration

Observing interactions among chemical species and microorganisms in the earths sub‐surface is a common task in the field of geology. Bioremediation experiments constitute one such class of interactions which focus on getting rid of pollutants through processes such as carbon sequestration. The main goal of scientists’ observations is to analyze the dynamics of the chemical reactions and understand how they collectively affect the carbon content of the soil. In our work, we extract the high‐level goals of geologists and propose a visual analytics solution which helps scientists in deriving insights about multivariate, temporal behavior of these chemical species. Specifically, our key contributions are the following: i) characterization of the domain‐specific goals and their translation to exploratory data analysis tasks, ii) developing an analytical abstraction in the form of perceptually motivated screen‐space metrics for bridging the gap between the tasks and the visualization, and iii) realization of the tasks and metrics in the form of VIMTEX, which is a set of coordinated multiple views for letting scientists observe multivariate, temporal relationships in the data. We provide several examples and case studies along with expert feedback for demonstrating the efficacy of our solution.