Larry Xu

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2017-97

May 12, 2017

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-97.pdf

Interactive data visualizations have become a powerful tool for exploratory data analysis, allowing users to quickly explore different subsets of their data and visually test hypotheses. While new design tools have made creating such visualizations easier, these tools often do not scale well to large datasets that do not fit on a user's machine. This becomes increasingly more common as data collection grows and storage costs become cheaper. In such cases, the visualization must request new data across a network for each interaction. When interactive latencies are introduced to a visualization, the user experience can deteriorate to the point of being inconsistent and unusable. To address these issues we present a new interaction design technique, <em>Chronicled Interactions</em>, which allows for creating asynchronous interactive visualizations that are resilient to latency. Through online user studies, we validate that users are able to complete visual analysis tasks quicker with this new design compared to a traditional interface under conditions of latency.

Advisors: Joseph M. Hellerstein

---

BibTeX citation:

@mastersthesis{Xu:EECS-2017-97,
    Author= {Xu, Larry},
    Title= {Supporting Asynchronous Interactive Visualization for Exploratory Data Analysis},
    School= {EECS Department, University of California, Berkeley},
    Year= {2017},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-97.html},
    Number= {UCB/EECS-2017-97},
    Abstract= {Interactive data visualizations have become a powerful tool for exploratory data analysis, allowing users to quickly explore different subsets of their data and visually test hypotheses. While new design tools have made creating such visualizations easier, these tools often do not scale well to large datasets that do not fit on a user's machine. This becomes increasingly more common as data collection grows and storage costs become cheaper. In such cases, the visualization must request new data across a network for each interaction. When interactive latencies are introduced to a visualization, the user experience can deteriorate to the point of being inconsistent and unusable. To address these issues we present a new interaction design technique, <em>Chronicled Interactions</em>, which allows for creating asynchronous interactive visualizations that are resilient to latency. Through online user studies, we validate that users are able to complete visual analysis tasks quicker with this new design compared to a traditional interface under conditions of latency.},
}

EndNote citation:

%0 Thesis
%A Xu, Larry 
%T Supporting Asynchronous Interactive Visualization for Exploratory Data Analysis
%I EECS Department, University of California, Berkeley
%D 2017
%8 May 12
%@ UCB/EECS-2017-97
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-97.html
%F Xu:EECS-2017-97