DataConf

DataConf aims at augmenting conference publications by providing seamless access to publications and authors' metadata, as well as other available services related to these publications.

Targeted users are conference attendees owning a recent, internet-enabled mobile device. This app does not require any installation process, since it is entirely Web-based.

• View the full PDF description of the app, accepted for the ESWC'2013 AI mashup challenge.
• View the poster representing DataConf and its interactions with linked DataSources (A1 paper format).

DataConf won 2nd place of the ESWC'2013 AI mashup challenge. Thanks to all the participants (and reviewers ) who liked our app and voted for us!

Follow DataConf on Twitter

The base DataConf URI is http://dataconf.liris.cnrs.fr/.

Available instances:

• ESWC'2013 : http://dataconf.liris.cnrs.fr/eswc2013/
• WWW'2012: http://dataconf.liris.cnrs.fr/www2012/
• ISWC'2012: http://dataconf.liris.cnrs.fr/iswc2012/

• WWW'2013 : coming soon

Previous version: WWW'2012 poster app

DataConf can provide the following functions:

• access to conference metadata (publications, authors, affiliations) and navigation among these metadata (following links on authors, publications and keywords) in Linked Data
• publication search by title, author or keywords
• integration of available Web services related to the conference venue, such as:
  • conference / session schedule, to find where and when a particular paper will be presented
  • descriptions and links to external resources related to a publication, such as PDF files or video presentations if any.

In addition, conference organizers can provide QRCodes that users can flash to gain direct access to a publication homepage.

Current version of the app is 1.2.

• Application core
  • Allows querying and integrating Linked Data sources. DataConf core is now configurable to allow deploying different instances for different conferences, with different datasources / features. Each datasource is processed in a separate component.
• Publication datasources
  • Main datasource (SWC): to browse any conference dataset in the Semantic Web Conference ontology format and view metadata of the conference publications, authors and organizations. Deployed for the ISWC'2012 and WWW'2012 conferences
  • DBLP: to access authors' other publications. Fully functional, but the endpoint seems to experiment latency issues (solved: switched to RKBExplorer)
  • DuckDuckGo!: to retrieve extra infos about authors.
  • Google AJAX API: to retrieve authors' homepages, since DuckDuckGo! was not precise enough
• Other datasources
  • SimpleSchedule: to view the conference, tracks and location schedules. SimpleSchedule has its own backoffice interface to allow conference chairs managing the conference events. Read more about this app at SimpleSchedule datasource for DataConf.
  • DataPaper: to access external resources about conference publications and authors. For each of them, DataPaper stores the URIs, description texts and format of these resources in a NoSQL database. A backoffice interface for enriching the DataPaper database is available as an unofficial WordPress plugin. Read more about DataPaper at http://liris.cnrs.fr/lionel.medini/wiki/doku.php?id=datapaper
  • Inference engine: to process the conference keyword ontology. It allows retrieving super and sub-keywords (and thus the publications they refer to), and recommending the user publications they did not yet see and the refer to keywords close to those the already explored. Even if it is local, the inference engine is processed as a datasource, so that it can be queried and integrated in the views in the same manner as external datasources. For this, we use the OWLReasoner inference engine.
• Extra functions
  • Graph-based interface
  • Web-based QR-Code scanner (requires a browser implementing the getUserMedia interface)

• Improvement of the reasoning process on the keyword ontology, using a custom OWLReasoner version and a direct injection of the reasoning database, instead of building it on the client side (this process took several minutes on a regular smartphone for a big conference dataset). We will also present a standard means of building a keyword ontology from a conference dataset, and linking it to the ACM computing classification categories.

Linked Data: DataConf instances can be deployed for any available conference dataset conforming to the Semantic Web Conference ontology, such as those available on the Semantic Web Dog Food endpoint. Publication metadata are enriched by DBLP and DuckDuckGo!.

Service integration: DataConf provides access to and easy integration of external Web services to the user interface, using a component-base approach that allows handling these services in the same way as Linked Data endpoints.

Mobile reasoning: for keyword search, during the navigation, users construct their own local ontology and classify it on client side using the OWLReasoner JS engine; keyword search consists in requesting this enriched ontology

Advanced Web technologies: DataConf takes advantage of recent W3C advances. For example, it makes wide use of the Web Storage (LocalStorage) specification and flashing a QR-Code can be done inside the webapp for browsers that support device API.

Since its previous versions, DataConf has been refactored for offering a lightweight and extensible architecture. We particularly focused on the following points:

Server independence: DataConf can be deployed on any kind of Web server, since all server-side contents are static and processing is done on client side.

Community-adopted solutions: In order to facilitate the understandability and reusability of the core DataConf architecture, we chose to rely on an existing JS framework. We adopted Backbone.js since it is sufficient for our needs (a single-page MVC application) and widely supported by the Web community.

Modularity: SPARQL endpoints and Web services are considered as “datasources” and processed by components declared in separate JS objects. The component interface contains three methods dedicated to asynchronous query construction, model callback and view callback. Our aim is that this simple model gets adopted by developers wishing to integrate new datasources to the interface.

Configurability: a simple JSON file is required to initiate a DataConf instance for a particular conference. It contains the conference global information, links to the files containing datasource manager components and layout templates, and the routes that associate user actions to datasource queries.

Conference Chairs can easily deploy a new DataConf instance for any conference with available metadata from a SPARQL endpoint. For this, they need to:

• unzip the sources (soon) available on the DataConf Github project on a Web server
• write a conference configuration file similar to this one.
• eventually, create new datasources components related to the specific tools used for managing the conference such as this one and use them in the configuration file. Do not hesitate to contact Lionel Médini to get more details about this process.

…Or contact us to have it hosted on our server.

This app is designed and developed by 1 researcher and 5 students of the University of Lyon (LIRIS laboratory / Université Lyon 1), Lyon, France.

• Lionel Médini: associate professor, LIRIS lab

• Florian Bacle: master degree student
• Benoît Durant de la Pastellière: master degree student
• Fiona Le Peutrec: master degree student
• Nicolas Armando: master degree student

• Hoang Duy Tan Nguyen: master degree student

• The DataConf logo was designed by Sandrine Atamaniuk.

DataConf by Lionel Médini is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported License.
Based on a work at poster.www2012.org.
Permissions beyond the scope of this license are available at http://liris.cnrs.fr/dataconf/doku.php?id=licence.