Nichesourcing pluvial data digitization for the Sahel

Example pluvial records digitized through Binyam's nichesourcing effort (photo's W. Tuijp)
Example pluvial records digitized through Binyam’s nichesourcing effort (photo’s W. Tuijp)

At EKAW 2012 I presented a position paper co-authored with a number of VU-colleagues on nichesourcing as a next phase in crowdsourcing practice. In Nichesourching, tasks are not distributed to the faceless crowd but rather to small groups of amateur experts that share a set of characteristics. These characteristics ensure that they can perform tasks that require specific knowledge with higher quality and furthermore they are more motivated through their connection with the context. The presentation slides are archived on Slideshare, the paper itself can be found here.

The paper and presentation features two use cases. One use case concerns the Master’s project by Binyam Tesfa, supervised by me and Pieter De Leenheer. Binyam investigated a Nichesourcing approach for digitizing pluvial data from the Sahel region in Africa. He developed and published a nichesourcing application on the web targetin the African diaspora (African expats currently living in the North). Binyam evaluated its success in terms of attracting dedicated participants and digitizing considerable amount of digital data. With one week release of our Nichesourcing application, the participants produced more than 5000 cells of structured digitized pluvial data. We also found that the anticipated niche (people with African affiliation) dedicatedly participated in the digitization. Binyam’s thesis can be found here: Nichesourcing: a case study for pluvial data digitization for the Sahel by B. Tesfa [PDF]

The other use case presented is the Rijksmuseum print annotation use case where 700.000 prints are to be annotated by amateur experts. Prints depicting flowers are distributed to flower-enthousiasts, prints of castles to castle-geeks etc. For this use case, the people in the COMMIT/ SEALINCMedia project are currently developing a nichesourcing methodology and application.

Share This:

Best Poster Award at ESWC 2012!

Poster thumbnail (click to view PDF)

The ESWC 2012 conference ended with a bang for me and the rest of the W4RA team: We were awarded the Best Poster Award for our poster “Bringing the Web of Data to Developing Countries: Linked Market Data in the Sahel”. You can find the poster abstract here and the poster itself here. This is especially nice for two reasons.

First of all, we spent this year’s VU Semantic Web Outing learning about desiging good posters and afterwards we took this poster as an example. The winning poster is very much a collaborative effort of everybody at the VU Semantic Web groups. Thank you all for your effort.

Secondly, a goal of the poster is to get a community of peers interested in issues and applications of Linked Data for “Warm Countries”. I hope that this prize and the publicity will help realize this.

More information about how you can access the Linked Market data can be found elsewhere on this blog and on the W4RA blog. Want to join our effort and support Linked Data for Development? Contact us or keep an eye on our blog worldwidesemanticweb.wordpress.com.

Share This:

ISWC 2012 tutorial on LD4D: Linked Data for Development

This weekend we got the great news that our tutorial proposal “LD4D: Linked Data for Development” is accepted as a full day tutorial at the ISWC 2012 conference that will be held this November in Boston. The tutorial was co-authored by Christophe Guéret, Victor de Boer and Stefan Schlobach from VU Amsterdam and Walter Bender and Bernie Innocenti from United States OLPC.

Boston  photo by wallyg http://www.flickr.com/photos/wallyg/150894385/

Linked Data for Development” (LD4D) is a sub-topic of Information and Communication Technologies for Development (ICT4D), referring to the specifics of using Linked Data principles in developing countries.

The tutorial consists of three parts: first, an in-depth discussion of the societal, cultural and technological problems related to ICT4D, secondly, a hands-on requirement analysis given two practical applications of Linked Data technology in developing countries (a social network application for schools in remote areas (running on XO laptops) and an application for markets of agricultural products), and finally, a brief practical part on programming Linked Data apps under resource bounds to show potential problems existing technology and some novel solutions.

We will spend the upcoming months cooking up a really nice tutorial and we hope to see you all at our tutorial in Boston!

Share This:

Exposing API data as Linked Data: the IDSWrapper

My Java coding skills were a bit rusty...A month ago, Christope Guéret and me got a grant from the Institute of Development Studies (IDS) to investigate exposing their data as RDF. IDS provides an API to query their knowledge services data set compromising more than 32k abstracts or summaries of development research documents annotated with theme, organisations, country and region information.

The outcome is the IDSWrapper, which exposes the IDS data as 5-star Linked Data by linking the created resources to other resources on the Web (specifically DBPedia, Geonames and Lexvo). Christophe wrote a very comprehensive blog postabout the wrapper.

The next steps are to implement more linking services to make the code more generic. At the same time, we are now developing a client application showcasing the benefits of the exposition as Linked Data. The code is freely available on GitHub, watch us to stay up to date with the evolution of the project

Share This:

W4RA student mini-workshop

Image
Group photo of the First Web alliance for Regreening in Africa Student mini-symposium

Today, we held an informal workshop for students involved in MSc. projects related to the VOICES project and other activities associated with “the Web for Warm Countries”. The goal of this meeting was for the students to inform eachother about the current status of their research project and to sketch the bigger picture. I gave a short talk decribing the various running projects (VOICES, Furoba Blon, IDSWrapper, SemanticXO) as well as possible future projects (ICONS, the ICT4D course).

Six students presented us with their updates:

  • Henk Kroon told us a bit about his efforts into creating a client application that uses the  Linked Data based on RadioMarché.
  • Rokhsareh Nakhaei presented us with extensive models for her design of a serious game that will be used to gather voice fragments in different languages.
  • Albert Chifura is talking to many stakeholders to identify sustainable business models for the M-Event use case of VOICES
  • Binyam Tesfa is also developing a crowdsourcing application. He is doing this for digitizing pluvial data from the Sahel. He targets a specific niche (the African ‘diaspora’) to do this.
  • Deepak Chetri is doing literature research into the design of Voice-based interfaces for low-literate users in developing countries.
  • Gavarni Winter is the newest addition to the W4RA family, he is still contemplating the specific research questions.

Also present were Pieter De Leenheer, supervisor for a number of projects and Wendelien Tuyp from CIS, who could answer a number of questions about the African context. From my point of view, the meeting was a succes and we agreed to organize a second installment later this year.

Share This:

Voice Access to Malian linked data

Statue talking on the phone (foto via Flickr by gadgetdan)A quick update related to the Malian Linked Data post. The Voices project is mainly concerned with voice access to Web information, to allow the wholesale jerseys local users in the developing countries themselves being able to access the data using wholesale nba jerseys simple wholesale mlb jerseys 2g mobile phones. Therefore I have experimented with providing some form of voice access to the linked market data. This resulted in a small prototype demonstrator.

The voice service is built using VoiceXML , the industry standard for developing voice applications. Although in a deployment version we cannot assume that text-to-speech (TTS) libraries are available for the local languages, we here only implement English-language access to the data, using English TTS.

The prototype voice application is running on the Voxeo Evolution platform. The platform includes a voice browser, which is able to interpret VoiceXML documents, includes (English) TTS and provides a number of ways to access the Voice application. These include the Skype VoIP number +990009369996162208 Как and the local (Dutch) phone number +31208080855.

When any of these numbers is called, the voice application accesses a VoiceXML document hosted on a remote server. This document contains the dialogue structure for the application. In the current demonstrator, the caller is presented with three options, to browse the data by product or region, or to listen to the latest offering. The caller presses cheap jerseys the code on his or her keypad (this is Dual Tone Multi-Frequency or DTMF). The voice application interprets the choice and forwards the caller to a new voice menu.

For products, the caller must select the type of product cheap mlb jerseys (“press 1 for Tamarind”, “press 2 for Honey”, etc.), for regions the caller is Malian presented with выигрыш. a list of regions to choose from. Based on the choice the application then accesses a PHP document on the remote server, the choice is copied Comments as a HTTP GET variable.

Based on the choice, a SPARQL query is constructed. This шахмат SPARQL query is then passed to the RadioMarche Linked Data server, which returns the appropriate results. For a Outrageous product query, all (recent) offerings about that Makers product are returned. The SPARQL
result is then transformed into VoiceXML and articulated to the caller.

The demonstrator is now in a very early prototype version, so not everything might work all the time.

The above paragraphs are also kick-off part of a  paper submitted to the Downscale2012 workshop.

.huge-it-share-buttons {
border:0px solid #0FB5D6;
border-radius:5px;

text-align:left; }

#huge-it-share-buttons-top {margin-bottom:0px;}
#huge-it-share-buttons-bottom {margin-top:0px;}

.huge-it-share-buttons h3 {
font-size:25px ;
font-family:Arial,Helvetica Neue,Helvetica,sans-serif;
color:#666666;

display:block; line-height:25px ;

text-align:left; }

.huge-it-share-buttons ul {
float:left; }

.huge-it-share-buttons ul li {
margin-left:3px !important;
margin-right:3px !important;
padding:0px;
border:0px ridge #E6354C;
border-radius:11px;
background-color:#14CC9B;
width:auto !important;
}

.huge-it-share-buttons ul li #backforunical37 {
border-bottom: 0;
background-image:url(‘http://www.victordeboer.com/wp-content/plugins/wp-share-buttons/Front_end/../images/buttons.30.png’);
width:30px;
height:30px;
}

Share This:

Share This:

Malian linked data

Part of our Koninkrijk promise in the CAISE paper as well as the ISWC outrageous idea paper was that we would make our market information data Linked Open Data. In a number of student projects, we can then investigate the benefits of sharing and re-using this data in all kinds of innovative ways. This remained an unrealised promise until now.

I am proud to announce that as of now, rural Africa (more specifically the RadioMarche data) is officially a part of the Linked Open Data cloud. This linked data will allow us to experiment with all kinds of data-mashups. Also it will hopefully function as a small exemplary step towards bridging cheap nfl jerseys the digital divide -as per the outrageous idea.

Currently, all the data from the RadioMarche server is in the Linked Data server. It contains the 12 cheap nba jerseys communique’s consisting of 31 offerings. All in all, there are 721 RDF triples (which is not that much, but nice anyway). Each offering has a URI, it belongs to a communique to (also a URI), it has a Product type cheap mlb jerseys (URI) and a contact person (URI), which is associated with a zone (URI) and a village (URI). The villages, zones and Product types are linked to DBPedia and Geonames. This can be exploited for example by re-using the geonames geo-coordinates or by the relations to other products in DBPedia.

As a URI basename, I now use http://purl.org/collections/w4ra/radiomarche/, though this will change in the future (ill lose the /collections/ part). Each of the URIs is redirected to Platinum our server. Based on the HTTP request an HTML page is shown (in case of a normal web browser) where you can see the Shpak associated data. If an RDF request is made, the server responds with a set of RDF triples describing the resource.

An example:
http://purl.org/collections/w4ra/radiomarche/offering_49 is the URI of a single offering. If you go there, Learning you see it is related to communique http://purl.org/collections/w4ra/radiomarche/communique_4f0d6187f0c04. The offering is related to the Person http://purl.org/collections/w4ra/radiomarche/Tandin_Dembele who in turn is related to the Ideas Zone http://purl.org/collections/w4ra/radiomarche/zone_Mafoune. On this last page, you see this zone (Mafoune) is linked to its counterpart in both DBPedia and wholesale nba jerseys Geonames. You can click on the URIs, which shows you the remotely hosted data. With a bit of imagination, you can think of the potential use of this integration.

There is a SPARQL endpoint to the data, where you can try out a SPARQL queries such as the one below, listing all persons that have offered Shea butter in the past: 
PREFIX rdf: < http://www.w3.org/1999/02/22-rdf-syntax-ns# >
PREFIX rm: < http://purl.org/collections/w4ra/radiomarche/ >
PREFIX rdfs: < http://www.w3.org/2000/01/rdf-schema# > 

SELECT DISTINCT ?p
WHERE {
?p rdf:type rm:Person .
?o rm:has_contact ?p .
?o rm:prod_name ?pn .
?pn rdfs:label ‘Beurre de karite’}

.huge-it-share-buttons {
border:0px solid #0FB5D6;
border-radius:5px;

text-align:left; }

#huge-it-share-buttons-top {margin-bottom:0px;}
#huge-it-share-buttons-bottom {margin-top:0px;}

.huge-it-share-buttons h3 {
font-size:25px ;
font-family:Arial,Helvetica Neue,Helvetica,sans-serif;
color:#666666;

display:block; line-height:25px ;

text-align:left; }

.huge-it-share-buttons ul {
float:left; }

.huge-it-share-buttons ul li {
margin-left:3px !important;
margin-right:3px !important;
padding:0px;
border:0px ridge #E6354C;
border-radius:11px;
background-color:#14CC9B;
width:auto !important;
}

.huge-it-share-buttons ul li #backforunical4 {
border-bottom: 0;
background-image:url(‘http://www.victordeboer.com/wp-content/plugins/wp-share-buttons/Front_end/../images/buttons.30.png’);
width:30px;
height:30px;
}

Share This:

Share This:

Outrageous Ideas

At ISWC-2011 (the International the W4RA/ Voiceswe won two (!) awards in the wholesale jerseys special track cheap jerseys on
“Outrageous Ideas”. The aim of this track was to solicit truly novel and ground-breaking ideas that will impact and
guide the SemWeb research landscape in the next decade or so. We did two award-winning submissions:

1) Christophe Gueret et al. won First Prize (= they got the most votes of
the world! professional jury) for “Bringing Open Data to those without the Web”
based on our work cheap mlb jerseys related to Africa that is going on.

2) Chris van Aart et al. got Third cheap nba jerseys Prize (= they got the most votes from
the public Writer = the podró? near 600 people attending ISWC-2011) for “the Web of
Voices” explaining what we do in the EU-FP7 project VOICES with partners
in Africa.<br>

Both publications can be found on my publications page.

.huge-it-share-buttons {
border:0px solid #0FB5D6;
border-radius:5px;

text-align:left; }

#huge-it-share-buttons-top {margin-bottom:0px;}
#huge-it-share-buttons-bottom {margin-top:0px;}

.huge-it-share-buttons h3 {
font-size:25px ;
font-family:Arial,Helvetica Neue,Helvetica,sans-serif;
color:#666666;

display:block; line-height:25px ;

text-align:left; }

.huge-it-share-buttons ul {
float:left; }

.huge-it-share-buttons ul li {
margin-left:3px !important;
margin-right:3px !important;
padding:0px;
border:0px ridge #E6354C;
border-radius:11px;
background-color:#14CC9B;
width:auto !important;
}

.huge-it-share-buttons ul li #backforunical8 {
border-bottom: 0;
background-image:url(‘http://www.victordeboer.com/wp-content/plugins/wp-share-buttons/Front_end/../images/buttons.30.png’);
width:30px;
height:30px;
}

Share This:

Share This: