Geography Reference
In-Depth Information
Atlases are particularly complex geovisualization environments, delivering
information on a large collection of themes (MacEachren et al.
2008
;O
Dea
et al.
2011
). As such, providing direct interaction options for all the underlying
information at once is out of question, and even organizing the interface categor-
ically in a way that allows access to all the main categories from the main interface
can easily cause information overload (Kramers
2008
; Bhowmick et al.
2008
). In
such information-rich environments, various design considerations should be taken
into account such as the layout design (e.g., where to place which control, in what
size, color, orientation or visual hierarchy based on semantic importance, labeling),
organization of the display elements (e.g., what to group together, whether to follow
conventions or not when utilizing the “screen estate”), and interaction design
(Galitz
2007
;¸¨ ltekin et al.
2009
).
An interaction design principle that is originally proposed for information
visualization displays which can be translated to GUI design is the three-level
“Shneiderman
'
s information seeking mantra”: overview—zoom-and-filter—details
on demand (Shneiderman
1996
). While “Shneiderman
'
s mantra” is a rather rea-
sonable starting point in organizing the elements of interaction, it appears that
evaluating or validating any design often (if not always) requires a user study.
Acknowledgement of the needs to place the user in the center of the design process
led to the term “user-centered design” (UCD) which is commonly used today (e.g.,
Fuhrmann and Pike
2005
; Holtzblatt et al.
2009
). UCD is a formalized approach
which ideally involves user feedback in all stages of iterative design: At the early
stages a feedback from the user should be obtained, integrated in the next stage of
design, and new user tests should be conducted to further improve the design until
the
'
solution is reached (e.g., Nivala et al.
2007
). However, it should be noted
that user-centered approaches appeared in cartography and geovisualization
designs even before the digital and interactive maps have become common (Bartz
1970
; Eastman
1977
via Olson
1979
). More recently, we see prototypical examples
in which users can personalize the display or the GUI as they wish (Balciunas
2013
)
or utilize Web 2.0 approaches such as map recommendations, user comments, tag
clouds and RSS feeds for online atlases (
¨
zerdem et al.
2013
).
The closest work to our project in the literature appears to be by Kramers (
2008
),
as this study features an evaluation of an atlas interface. Kramers (
2008
) highlights
the importance of understanding the users, individual and group differences among
users, their
right
'
'
behavior, the task and the context of use. Presenting how
user-centered design improved the user performance over time with the Atlas of
Canada, Kramers (
2008
) focuses on the evolution of a particular design. Contrary to
this, in our project we evaluate five alternatives with a specific age group. In this
project, we draw concepts from earlier literature as well as from modern interface
examples that may be well-known by digital natives. Using five alternative GUI
designs in which we allow limited interaction, we experiment with layout density
(linked to information overload concept), visually grouping the elements based on
their function (or semantic closeness) on the interface and some standard interac-
tion paradigms (linked to information seeking mantra).
map use
'
'
