See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262164528

Democratizing mobile app development for disaster management

Conference Paper · August 2013

DOI: 10.1145/2516911.2516915

CITATIONS READS
25 996

6 authors, including:

Oshani Seneviratne Carlos Castillo

Rensselaer Polytechnic Institute University Pompeu Fabra
42 PUBLICATIONS   299 CITATIONS    264 PUBLICATIONS   12,275 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Assessing the Impact of Music Recommendation Diversity View project

Crisis Management View project

All content following this page was uploaded by Carlos Castillo on 22 June 2014.

The user has requested enhancement of the downloaded file.

 Democratizing Mobile App Development for Disaster Management
Fuming Shih, Oshani Seneviratne, Daniela Miao, Ilaria Liccardi, and Lalana Kagal
Massachusetts Institute of Technology
Cambridge, USA

Evan Patton Carlos Castillo, Patrick Meier

Rensselaer Polytechnic Institute Qatar Computing Research Institute
Troy, New York Doha, Qatar

Abstract technical volunteers have developed useful mobile applica-

tions for disaster reliefs. However, each organization usually
Smartphones are being used for a wide range of has its own application that creates or consumes data stored
activities including messaging, social networking, in standalone databases, or even worse, in spreadsheets. This
calendar and contact management as well as loca- means most of these applications are built in silos without ex-
tion and context-aware applications. The ubiquity ploiting the potential of being “interlinked” with other data
of handheld computing technology has been found collected from other organizations, or even the government.
to be especially useful in disaster management and The lack of expertise and the cost for building mobile appli-
relief operations. Our focus is to enable developers cations cause relief workers to turn to more manual steps for
to quickly deploy applications that take advantage merging various reports from volunteers. Consequently, the
of key sources that are fundamental for today’s net- increasingly unorganized and scattered information become
worked citizens, including Twitter feeds, Facebook noise in the system and can sometimes slowdown the deci-
posts, current news releases, and government data. sion making process. We envision a framework that enables
These applications will also have the capability of non-technical application developers to quickly build and de-
empowering citizens involved in crisis situations to ploy applications that easily reuse existing and crowdsourced
contribute via crowdsourcing, and to communicate structured information sources.
up-to-date information to others. We will lever- One major challenge in the afore-mentioned example is
age several technologies to develop this application how to effectively integrate data generated by multiple par-
framework, namely (i) Linked Data principles for ties, including the “crowd”, relief organizations, and govern-
structured data, (ii) existing data sources and on- ment agencies. We address this challenge by using Linked
tologies for disaster management, and (iii) App In- Data principles2 to enable the reuse, extension and integration
ventor, which is a mobile application development of heterogeneous structured data from distributed sources.
framework for non-programmers. In this paper, we Linked Data is a set of design principles proposed by the
describe our motivating use cases, our architecture, World Wide Web Consortium (W3C) for using Web tech-
and our prototype implementation. nologies in order to support the distributed development of
structured information such that it can be easily and automat-
ically combined. However, consuming and generating Linked
1 Introduction Data is difficult as pointed out in [Scharffe et al., 2012], es-
Smartphones are quickly becoming the primary computing pecially on mobile devices. This implies that we need to also
and communication platform for people’s daily tasks. With focus on reducing the barrier to the adoption of Linked Data
the rise of social networks, “netizens” are now comfortable technologies on smartphones.
with frequently updating their social profiles with their cur- This paper describes our ongoing work of extending the
rent activities and/or locations. This new source of informa- App Inventor3 platform with Linked Data technologies. App
tion, social signals from microblog platforms, has been found Inventor is an open-source project that provides a block-
to be especially useful in disaster management and relief op- based interface to visually create Android apps [Wolber et
erations [Fajardo and Oppus, 2010]. For example, during al., 2011]. We are developing “Linked Data” components for
the Beijing flash floods in July 2012, people took to Twit- App Inventor that will allow app developers to easily build ap-
ter to provide information about the flooded areas and col- plications that explore and consume Linked Data datasets as
laboratively developed a live crisis map of the floods impact well as publish structured data directly to remote Linked Data
using Google Maps 1 . By integrating crowdsourced infor- repositories. Our App Inventor components will focus on (i)
mation with Geographic Information Systems (GIS) data or embedding semantic concepts directly into the process of mo-
other open datasets released by the local government, some
2
http://www.w3.org/DesignIssues/LinkedData.html
1 3
http://iRevolution.net/2012/08/01/crisis-map-beijing-floods/ http://appinventor.mit.edu/
 9:40 9:40
QCRI APP QCRI APP 9:40 9:40
WATER QCRI APP
Donate FOOD

WATER

MEDICINE
Request
HOSPITAL
9:40
QCRI APP SHELTER
Recent Actions
CLOTHES

Welcome to the QCRI app.

Log in using your twitter credentials (a) (b)
9:40 9:40
WATER
QCRI APP
Recent Actions

username Name
Tag
WATER
I can donate water 3L.

9:23 AM
zip code
Edit Remove

password WATER
if blank we will use your GPS Location
Description
I can donate water 5L.
Description
10:13 AM Edit Remove

FOOD
I can donate food.
I am looking for water, please help.
Log in 10:13 AM Edit Remove
Submit
9:23 AM reply
Submit

(c) (d) (e) (f)

Figure 1: Two Semantic Web enabled mobile apps for disaster response situations

9:40
bile app-building while hiding most of the operational details Microblog  
WATER
from novice developers, and (ii) the integration of structured Analysis  
information from microblog platforms, crowdsourced and ex-
isting open data.
Data  service   Data  
2 Motivating scenarios endpoints   Cura7on  
We describe two applications that can potentially be used by
people during a crisis situation: (1) App Donate-N-Request
will be used to match requests and donations during a disaster
scenario, and (2) App WeReport will be used to report scenes Developers Crowd with LD-Apps Crowd
and information related to the disaster. These apps will have
9:40
the ability to be integrated with different public datasets using
There are 8 people looking for water in your proximity Figure 2: Architecture of LinkedWATER
Data enabled crowdsourcing
linked data. and application development platform I am looking for water, please help

9:23 AM

2.1 Donate-N-Request App I am in your proximity, do you still need water?

This app is inspired by MatchApp4 , where requests for re- some can be dangerous for the pedestrians. With our WeRe-
10:23 AM

sources are met with the availability of those resources. port app, he could take either a picture or a video of the scene.
Consider this resource match-making scenario: Alice lives He decides to take a picture of the road, by adding a “tag” and
in New York City. After Hurricane Sandy hits, 9:40 Alice wants a “description” to the image as shown in Figure 1(f).
to help people in need. QCRIBob
APP has been affected by the crisis.

Both Bob, the requestor, and Alice, the donator, can place the 3 Prototype Implementation
donation and request for Take a picture or a video
items within our app as shown in
In order to implement the apps described in Section 2, and
Figure 1(a) & (b). They can also view/edit/remove recent list
gather the necessary data items, we explored the use of exist-
of their activities as shown in Figure 1(c). Alice chooses to
ing linked data vocabularies and tied them into App Inventor
donate water, so she checks up all the people within a 3 miles
as illustrated in Figure 2.
radius of her location who are in need of water on a map as
shown in Figure 1(e). Bob had previously requested water by 3.1 Linked Data
using our app (Figure 1(d)). Alice selects Bob from the list
of people in need. Alice and Bob can message directly with Disaster management activities such co-ordinating relief op-
each other to arrange the exchange. erations, setting up shelters, reporting structural damages, and
overseeing volunteers generate a lot of data. These need to be
2.2 WeReport App integrated and be interoperable with data provided by vari-
On his way to the grocery store, after a severe hurricane, Joe ous organizations to be useful in decision making processes
noticed many parts around his community are in bad condi- during disasters. Previous research [Babitski et al., 2011] has
tions. For example, fallen trees are blocking the road and demonstrated how the use of ontologies and Semantic tech-
nologies can be advantageous in developing applications for
4 disaster management scenarios. We identified two vocabu-
http://iRevolution.net/2013/02/27/matchapp-disaster-response-
app/ laries that are useful in modeling such data as Linked Data.
They have varying degrees of support for some of the disaster
concepts, and we used terms from both the vocabularies in
our implementation. We discuss some shortcomings of these
ontologies in Section 4.
Humanitarian eXchange Language Situation and
Response Standard (HXL)
HXL5 was developed through an initiative by the United Na-
tions High Commissioner for Refugees (UNHCR) after in-
specting a broad range of systems in use with different hu- (a) (b)
manitarian organizations. The schema is available in five
main sections, namely: geolocation (information on the lo-
cations affected), humanitarian profile (information on the Figure 3: Interface of the App Inventor platform
populations affected), metadata (who is collecting the data,
date/time the data was collected, etc), response (information
on the organization responding), and situation (information
related to the emergency).
Management Of A Crisis Vocabulary (MOAC)
MOAC6 was created after the earthquake in Haiti in 2010.
Much of the focus in developing this vocabulary was to miti- Figure 4: App Inventor autocompletion of concepts and prop-
gate the inconsistencies between aggregated data and require- erties powered by SPARQL queries over cached semantic
ments of relief organizations. This was the first Linked Data web ontologies
vocabulary that had the objective of facilitating non-experts to
provide data in Linked Data formats through crowd sourcing
mechanisms such as Google Map Maker, Open Street Map, Exploring Linked Data
Twitter and Ushahidi. MOAC defines three main sections that One challenge of using Linked Data is finding details of an
relate to relief agency standards. ontology that describes a dataset. App developers should not
be burdened with knowing all the details of an ontology. For
3.2 App Inventor
example, when building a crisis response app, a developer
App Inventor provides a Web-based development platform may begin with a general concept such as Shelter and let
for novices with no prior programming experience to create the system provide more relevant information. Our extension
mobile applications easily. A user can drag and drop graph- provides an autocompletion capability in the designer inter-
ical objects on the interface editor (Figure 3(a)) to design an face to assist developers in selecting the appropriate ontolog-
app or implement behaviors with blocks as shown in Figure ical term (see Figure 4). We provide a service within the App
3(b). App Inventor also provides many high-level compo- Inventor that loads ontologies, indexes labels for classes and
nents for using smartphone resources as well as communicat- properties, and answers queries to support this autocomple-
ing with external web resources. For example, it has com- tion widget.
ponents for reading the GPS location sensor, taking photos
with the camera, receiving SMS texts, and sending Twitter
Consuming Linked Data with Semantic Web Component
messages. By hiding most of the technical details, App In-
ventor gives its users the benefits to focus on designing app The Semantic Web Component converts semantic forms
behaviors rather than lower-level debugging. However, App into RDF graphs, executes and process results of SPARQL
Inventor currently has little support for integrating application queries, and saves and loads the content of ontologies. For
data with backend data-stores, for example, to save applica- example, the app developer can execute a SPARQL query to
tion data as Linked Data or integrate with other Linked Data fetch information about nearby shelters based on the user’s
resources. current GPS location and show the results on a map.
3.3 Linked Data Extension of App Inventor Publishing Linked Data with Semantic Form
In order to enable developers using the App Inventor platform
to interact with Linked Data sources, we provide an extended The Semantic Form is a layout component that can be used
version of App Inventor with additional components and user by developers to identify a collection of fields that should be
interface elements powered by a modified version of the Jena applied to a particular ontological concept. Forms can also
framework [Carroll et al., 2004]. In this section we outline assign an auto-generated subject URI based on one or more
how we expose Linked Data concepts to application develop- fields to create new Linked Data instances. For example, a de-
ers in the App Inventor the new components that can be used veloper can have two textfields for the name and the descrip-
by developers to consume and generate Linked Data. tion of a shelter. Semantic Form will automatically generate
Linked Data triples if those fields are tagged with ontological
5
http://hxl.humanitarianresponse.info/ns/index.html terms like the properties shelterName and shelterDescription
6
http://observedchange.com/moac/ns/ from the MOAC ontology.
 4 Discussion Using linked data within a mobile environment has proved
Existing disaster management vocabularies have support for useful and challenging in other research. David et al. [2010]
many disaster management scenarios. But there are some propose a general framework in order to introduce Linked
shortcomings in current vocabularies. HXL defines admin- Data sources as device content in the Android platform;
istrative units that may accept the donations, and the af- D’Aquin et al. [2010] consume selected data sources and
fected populations that would receive those donations. How- create new social link between people in the author’s univer-
ever there is no support to specify what the donation is, i.e. sity; Razzak et al. [2010] utilize government data to sup-
whether it is money, clothes, food, etc, as well as who is do- port tourism within the author’s city. However in all these
nating it and other logistics related to the donation. Similarly, approaches the application is domain specific and cannot be
MOAC has support for specifying the needs of a specific shel- extended to be used with different data sources. Extending
ter and who is handling what in the WhoWhatWhere (W3) the App Inventor framework with Linked data will allow de-
section of the vocabulary. However, there is nothing to de- velopers to be able to reuse available linked data sources and
scribe data relating to donations. Another shortcoming of to tailor and extend apps to a variety of scenarios.
these two vocabularies is the lack of support for situational
awareness data. As an example, if someone is reporting struc- 5 Summary
tural damage to a bridge that might be affecting a population, In this paper we have demonstrated the effective use of
there is no way to link the bridge data to MOAC and/or HXL. Linked Data within mobile applications with a special focus
We are in the process of formulating a vocabulary that will on disaster management. As organizations continue to expose
tie in these disaster vocabularies with other related linked data their data as Linked Open Data, our platform will be able
vocabularies, e.g. Geonames, Semantically Interlinked On- to help developers using the Linked Data App Inventor plat-
line Communities (SIOC)7 , etc. Geonames can be used to form to come up with useful applications quickly and without
model geo-location information, and the related population much effort to help those in need during disasters.
information. SIOC can be used to tie in information available
in microblog posts from tweets and Facebook posts after a References
disaster such as zones that people need to evacuate from, new
[Babitski et al., 2011] Grigori Babitski, Simon Bergweiler, Olaf
developments in a disaster (e.g. the potential sighting of a Grebner, Daniel Oberle, Heiko Paulheim, and Florian Probst.
looting), donation pledges along with requests for donations, SoKNOS - using semantic technologies in disaster management
etc. software. In Grigoris Antoniou, Marko Grobelnik, Elena Sim-
With the proliferation of linked datasets, developers us- perl, Bijan Parsia, Dimitris Plexousakis, Pieter Leenheer, and Jeff
ing App Inventor may want to use these data directly via Pan, editors, The Semanic Web: Research and Applications, Lec-
App Inventor components without having to write SPARQL ture Notes in Computer Science, pages 183–197. 2011.
queries themselves. In order to facilitate this feature, we [Carroll et al., 2004] Jeremy J. Carroll, Ian Dickinson, Chris
are in the process of developing “data wrapper” components. Dollin, Dave Reynolds, Andy Seaborne, and Kevin Wilkinson.
These components have built-in SPARQL queries to access Jena: Implementing the semantic web recommendations. In Pro-
frequently accessed data, with extended functionality to plug- ceedings of the 13th international World Wide Web conference,
in custom SPARQL queries to get at more specific data items. pages 74–83, New York, NY, USA, 2004. ACM.
With this functionality, we hope not to clutter the App Inven- [d’Aquin et al., 2010] Mathieu d’Aquin, Fouad Zablith, and Enrico
tor component interface with data items that a majority of app Motta. wayou - linked data-based social location tracking in a
inventor developers may not use while providing the flexibil- large, distributed organisation. 2010.
ity to retrieve any kind of data the developer desires. As an [David and Euzenat, 2010] Jérôme David and Jérôme Euzenat.
example, consider a data wrapper component for the NYC Linked data from your pocket: The android rdfcontentprovider.
demographics statistics datasets8 . This dataset includes all In ISWC 2010, 2010.
the population statistics such as number of males and females [Fajardo and Oppus, 2010] Jovilyn Therese B. Fajardo and Car-
as well as number of people based on their ethnic categoriza- los M. Oppus. A mobile disaster management system using the
tion in a given zip code area. This ethnic categorization in- android technology. WSEAS TRANSACTIONS on COMMUNI-
cludes generic categories such as white, black, hispanic, asian CATIONS, 9, June 2010.
in addition to very specific categories such as pacific islander, [Razzak et al., 2010] Faisal Razzak, Dario Bonino, and Fulvio
asian non-hispanic, white non-hispanic, white hispanic, black Corno. Mobile interaction with smart environments through
non-hispanic, black hispanic, etc. The data wrapper would linked data. In IEEE International Conference on Systems Man.
expose the gender based information and the general ethnici- and Cybernetics, 2010.
ties in the interface with built-in SPARQL queries that could [Scharffe et al., 2012] François Scharffe, Ghislain Atemezing,
just be dragged and dropped on to the interface, whereas very Raphaël Troncy, Fabien Gandon, Serena Villata, Bénédicte
specific queries to obtain information about a particular eth- Bucher, Fayçal Hamdi, Laurent Bihanic, Gabriel Képéklian,
nicity (for e.g. black hispanic) could be written by the app Franck Cotton, et al. Enabling linked-data publication with the
inventor developer and plugged into the component. datalift platform. 2012.
7
[Wolber et al., 2011] David Wolber, Harold Abelson, Ellen Sper-
http://sioc-project.org tus, and Liz Looney. App Inventor: Create Your Own Android
8
https://nycopendata.socrata.com/Statistics/Demographic- Apps. O’Reilly Media, 2011.
Statistics-By-Zip-Code

View publication stats