Database Reference
In-Depth Information
photos across the globe, along with the ability to select geographic regions for
annotation with names and additional data (Fig. 2). View sharing is supported
through automatically updating URLs. As the view is panned or zoomed, the
current URL updates dynamically to reflect the current zoom level and latitude
and longitude values. Pointing is supported through annotations. Users can draw
rectangular and polygonal annotations, which scale appropriately as the map is
zoomed. To avoid clutter, annotations are filtered as the view is zoomed; the
viewer does not see annotations that are too small to be legible or so large they
engulf the entire display, improving the scalability of the system.
Wikimapia supports conversation using an embedded discussion technique.
Each annotation is a link to editable text. Descriptive text about a geographic
region can then be edited by anyone, similar to articles on Wikipedia. Discussion
also occurs through voting. When annotations are new, users can vote on whether
they agree or disagree with the annotation. Annotations that are voted down
are removed from the system. For instance, the small town of Yelapa, Mexico
is located on an inlet in a bay near Puerto Vallarta. However, the bay has a
number of inlets very close together. As a result, multiple conflicting annotations
for Yelapa appeared. Through voting, the incorrect regions were discarded and
the correct annotation was preserved.
sense.us: Social Data Analysis of U.S. Census Data:
Sense.us is a proto-
type web application for social visual data analysis [46]. The site provides inter-
active visualizations of 150 years of United States census data, including stacked
timelines, choropleth maps, and population pyramids. With a URL-based book-
marking mechanism, it supports collaboration through doubly-linked discussion,
graphical annotations, bookmark trails, and searchable comment listings.
Discussion occurs via a doubly-linked conversation model. Searchable com-
ment listings provide links back into the visualization, while navigating in a visu-
alization automatically causes related comments to be retrieved for the current
view. By tying commentary directly to specific view states, comments become
less ambiguous, enabling remarks such as “that spike” or “on the left” to be more
easily understood. Pointing occurs through freeform graphical annotations and
view sharing is facilitated by URLs that automatically update as users navigate
the visualizations. Sense.us also allows users to collect view links in a “bookmark
trail” of view thumbnails. Users can then drag-and-drop view thumbnails from
the trail into a comment text field, thereby adding a hyperlink to the saved view
within the comment. In this way, users can provide pointers to related views and
create tours through the data.
Studies of the sense.us system revealed interesting patterns of social data
analysis. Users would make observations about the data, often coupled with
questions and hypotheses about the data. These comments often attracted fur-
ther discussion. For example, within a visualization of the U.S. labor force over
time, a spike and then decline in the number of dentists prompted discussion
ranging from the fluoridation of water to stratification within the dentistry pro-
fession, with a rise in the number of hygienists corresponding to the decline of
dentists. There was also an interesting interaction between data analysis and
Search WWH ::
Custom Search