(Shirky 2009, Laituri and Kodrich 2008). Crowdsourcing thus
thrives only when sufficient communication channels are
available. Applications aiming to leverage crowdsourcing must
supply the crowd with sufficient and efficient means of
communication.
Crowdsourcing strengths and weaknesses: crowdsourcing is
slowly finding its way into disaster management (Lucaszczyk
2011). Amongst the numerous crowdsourcing virtues, the
following are recognised to be valuable for disaster management.
Speed: crowdsourcing initiatives need little effort to materialize
and start cooperating An Ushahidi instance can be up and
running in two hours. The OpenStreet mappers have produced
completely new and highly detailed maps of Haiti in the course
of days (Harvard Humanitarian Effort 2011). Traditional
organizations tend to be slower in their response (Shirky 2009).
Up-to-date data: crowdsourced data can be collected at a
tremendous pace and kept fresh due to the "many eyes
watching' principle. Information is shared easily through
Ushahidi and Shahana, but also through blogs, Twitter,
Facebook, etc. while geographical information can be distributed
through platforms such as GeoNode and other OGC products.
Wide knowledge pool: as discussed above, crowdsourcing
initiatives are characterized by a widely diverse group of
contributing volunteers.
Momentum: Due to their openness (crowdsourcing initiatives
use the web to communicate and open source tools to
collaborate) crowdsourcing initiatives gain momentum faster and
keep it going for longer than closed organizations.
Continuity: a substantial part of volunteered (geographical)
information or disaster management software is the product of
free time activity and, to a lesser degree as a by-product of
commercial processes. As such, volunteers are constantly
working on, and are surrounded by the information and tools
that they later deploy and use during a disaster management
operation. The so created continuity ensures an efficient and
effective deployment and usage of the technologies. Although
official disaster management agencies organize training sessions,
disaster management is often one of their many tasks and is
certainly not a day -to-day experience.
The biggest threat to acceptance of crowdsourcing results, and
especially data, has always been the question of reliability and
robustness (accuracy when talking about data). Flanagin and
Metzger (2008) discuss these issues in terms of believability or
credibility-as-perception. The degree of believability is
determined by trustworthiness and expertise. Goodchild (2007)
and Schmitz et al. (2008) note that volunteered efforts can be
trustworthy even when not produced by experts by relying on
the collective "wisdom" of the crowd to detect and correct
inaccurate information entries, keep the data set up -to-date and
"defend" it from vandals and bugs.
WebGIS for crowdsourced disaster management: Traditional
GI systems are holistic, heavy weight solutions i.e. a single GI
system is designed to solve many a problem. GI systems need
powerful desktop computers, constraining GIS experts to a
desk. GIS 'in the field' e.g. in the hands of first responders and
volunteers has not seen a lot of practical application.
The dynamic nature of modern web pages and applications,
made possible by Web 2.0 technologies has started to move GIS
away from desktop machines. These technologies make it casy
to connect mobile applications to GIS servers through the
Internet in an interactive manner. Mobile devices have become
gateways to powerful servers that house geographical data and
perform complex analyses. Such mobile and lightweight GI
systems are called WebGIS. From the user's point of view, a
WebGIS is capable of performing the standard GIS operations,
but now users can take that functionality with them on the road.
Although several web based emergency management systems are
available, the spatial analyses supported by them are limited.
Acuna et al. (2010) have surveyed these systems and based on
their findings and performed literature studies, have identified
several design patterns they deem necessary for disaster
management web applications. These are listed as
e. Awareness for First Responders: fast and dynamic access
to information regarding the emergency at hand.
e. Collective Memory / (Temporal) Data Archives
e Tabular Information Presentation
e Map-based Navigation and Information presentation
e. Data authoring: mechanisms for attaching author and
source information to data items.
e Display of up-to-date Data
e Mechanism for Direct Data Manipulation
° Style Sheets for Multiple Media Types
° Hand-held Devices
Neis et al. (2010) present an emergency routing application
based on OpenStreetMap data and Open Location Services
Route Service. Their application supports shortest path
analysis on a damaged infrastructure network by allowing users
to mark blocked roads by drawing lines. The application is built
inside Ushahidi and thus leverages its crowdsourcing powers.
Drawbacks of this implementation are its interface, speed and a
lack of mobile client.
3. TECHNOLOGY
Based on the crowdsourcing and disaster management design
patterns discussed in sections 2.land 2.2, a small WebGIS has
been implemented that aims to enhance and automate the
disaster management activity of route finding with information
gathered through crowdsourcing.
The built application is a client-server configuration. The client
side runs on HTML/JavaScript and communicates through
asyncrhonous Javascript and XML or AJAX with a RESTful
server. The JavaScript library jQuery^ is used to implement
AJAX. REST is an abbreviation of REpresentational Ste
Transfer, a "software architecture for distributed hypermedia
systems such as the World Wide Web" (Pautasso et al. 2008). In
this architecture, every resource is stored on a server and has à
* http://www jquery .com
448
unique
as a
manip
Obtair
sendin
proces
resour
interck
plain t
REST
app lic:
by aut
and by
can the
makes
librarie
manipt
simp lic
docum
applica
require
Data is
Tables.
known,
momen
cloud-b
designe
and us.
2010b).
Tables
people
editors.
be visu:
à map.
as KMI
The us
applicat
is hoste:
4.
Geograp
data gat
presenta
these by
mechanis
for disas
Non-exp
desktop
———
htt]
6
WW
"htt
* hr i
