Entry Name: "UKON DBVIS-MC2"
VAST 2013 Challenge
Mini-Challenge 2: Situation Awareness Display Design
Adaptive User-Aware Dashboard Design
Team Members:
Fabian Fischer, University of Konstanz, Fabian.Fischer@uni-konstanz.de
Dominik Jäckle,
University of Konstanz, Dominik.Jaeckle@uni-konstanz.de
Dominik Sacha,
University of Konstanz, Dominik.Sacha@uni-konstanz.de
Florian Stoffel, University of Konstanz, Florian.Stoffel@uni-konstanz.de
Student Team: NO
(PhD Students)
Software Used: D3,
Firefox, Camtasia, Inkscape,
Gimp, OmniGraffle, Greenshot
May we post
your submission in the Visual Analytics Benchmark Repository after VAST Challenge
2013 is complete? YES
Video:
High-Resolution Image:
click here to open the image (PDF Format)
Storyboards: None
Description of Your Design:
1.1 Overall Concepts
Our concept has three important key concepts incorporated within its
design: The first one is the (1) User-Aware Adaptiveness
(UAA) of the dashboard, which utilizes a person tracking system to capture
the human’s awareness, to automatically switch between different widget
designs. This novel approach automatically adapts the dashboard according to
the attentiveness of the analyst to enhance real-time interaction in control
room situations with limited number of analysts. By constantly focusing the
dashboard, it will switch to the “Live Indicator Mode”, that shows the current
situation in a simplified form (e.g., only single number indicator, instead of
full line chart) to intuitively show the most important aspects. If the analyst
does not focus the dashboard the view will be switched to “Differential Mode”,
which keeps tracks of all occurring changes since the analyst last focused the
dashboard. This view is less intuitive and conveys more information in order to
boost the awareness of possibly missed occurred events in the past time the
analyst didn’t monitor the dashboard.
We make use of our novel approach, called (2) Shared Collaborative
Queues (SCQ), which makes it possible to save the current dashboard view to
a central collaborative queue. This queue can be accessed by individual
analysts - e.g. through tabletops - to drill down into observed suspicious
situations.
The last concept consists of different (3) Novel Widget Charts
that are included in our design proposal. To focus on different aspects of the monitored data and to ultimately gain
situational awareness, we propose the following widgets:
a) Indicator Overview Widget
The Overview Indicator Widget is meant to provide an overview over
different status indicators. This widget contains the most important network
metrics: health, security and performance. Each of the displayed values is
visualized with two nested rectangles: The outer one determines the outer
border of the visualization, the inner one is scaled according to the value of
the displayed condition. It is also filled with color, where black indicates
not having data at all, and red indicates the highest value (critical
condition).
One part of the Indicator Overview Widget showing the general Security
indicator having an alarming high value (large inner rectangle, strong red
color).
b) Bullet-Spark Chart Widget
The bullet graph is a chart type originally developed by Stephen
Few, which proved to be very effective in dashboard designs (Stephen Few, [Source]).
We make use of this graph to represent single number measures in our
dashboard as seen in the following figure.
We also introduce a modified version, which embeds sparklines
into the bullet chart. These are only visible when the dashboard is switched to
differential mode, as it overloads the analyst with information, which is not
always desirable. However, it gives a good overview of past changes of the
respective measures, when the analyst was not in front of the dashboard
application to quickly recap the last hours, as represented in the following
image.
c) Maptrogram Widget
The Maptrogram widget provides a top level
hierarchy overview of the globally-operating company network. The circles
represent the most active subnets and its size the amount of devices attaches
to the subnet: the bigger the circle, the bigger the subnet. Furthermore, the
circles are placed according to the location of the subnet. The color indicates
the aggregated status of the network. The arcs enclosing theses root-circles
correspond to the incoming and outgoing subnet flows. In addition, the upper
levels of the subnet hierarchies are added to the root-circles. All leafs of
each single subnet tree are positioned on a common line above and below the
map. To identify the source of incidents, an icon - added to the corresponding
leaf - reveals the dominating incident type (performance, health, or security)
within the subnet. In order to inspect the network hierarchy, the nodes are
linked with other widgets (e.g. the treemap).
Besides, the nodes are interactively expandable and can change their layout
according to the representation needed by the analyst.
d) Netrium Widget
The so called Netrium is the combination of
the words “network” and “aquarium” allowing the surveillance of networks over
time at a glance. An advantage of the netrium is the
combination of overview and detail. It offers an easy understandable overview
but adheres to several details: The x-axis maps the security status and the
y-axis maps the health status of the corresponding subnet. The analyst can
easily identify changes by keeping a focus on the quadrants. Moreover, the
moving circles correspond to a certain subnet mapping the size to the subnet
size. The pie chart contained by each circle displays the amount of connections
in the whole subnet. One of the most important features of the Netrium to notice major changes over time is the visualized
path of each circle. Also, a sparkline in the
background denotes the overall development of the network. To reduce clutter,
only countries above a certain threshold are visualized.
e) Treemap
The Treemap Widget displays the three most important system status
indicators and an aggregated measure for different subsidiaries of the company.
The number of systems in the network is represented by the cell size and hence
reflects its importance. The cells background color reflects the overall subnet
status on a color range from green (normal) to red (critical condition). Each
of the Health, Security, and Performance indicators are shown via a bar, where
the status is mapped to the height (low: Critical, high: Normal). To be able to
easily recognize the status of each of the bars, the background is colored in
red (Critical), yellow (Routine) and green (Normal). If the cell is not big
enough to hold the bar chart, the background color can be interpreted as the
current system state and the bar chart is left out. To compare the current
system state with the past, small indicators of the overall state are shown
below the cell name. Each indicator is colored with the same color scheme as
the cell background.