Over 45 participants attended the first hangout organized by BigDataEurope’s Transport domain in 2016, on 27 June, giving a technical overview of the Transport pilot that has been in the focus of the activities since the beginning of the year.
Simon Scerri from Fraunhofer IAIS started off by giving a brief introduction to the BigDataEurope activities and the Societal Challenge of Transport.
Next, Hajira Jabeen from the University of Bonn introduced the general architecture of the big data platform the project is set to build. The presentation covered the initial architecture and changes that have been made along the way, the components and also a gave a quick glimpse at the user interface. Have a look at her presentation below.
The second part of the hangout focused on the Transport pilot, taking place in Thessaloniki under the co-ordination of CERTH-HIT. Josep Maria Salanova gave an overview of the current ITS layout in the city, and the benefits that a big data platform has on enhancing the data analysis.
The final speaker, Luigi Selmi, introduced the technical background and details of the Transport pilot platform. He ave a detailed overview of the platform architecture and components and the further work to be done.
The final part of the 1 hour session was dedicated to answering the many questions participants raised during the presentations. Many questions inquired about the quality of the data and hence the quality of the service provided to users. Luigi Selmi replied stating that the quality is ensure by the data sources, which come directly from the devices therefore human error is eliminated. In addition in terms of the meta data, instead of using legacy meta data, ontologies will be used that improve the re-use of the data. Josep Maria answered that before the data sets they were relying on models which made accurate predictions challenging. However with the use of real data sets the prediction accuracy was significantly improved. Also, a thorough analysis and verification process is in place prior to data publication. Although this hinders incident recognition it could be balanced with pattern recognition work is now also focusing on improving the service in the case of incidents (when users need the service most) with the help of pattern recognition.
To view the full recording, click through.