Unister and Brox hosted the 3rd GeoKnow Plenary Meeting in the beautiful city of Leipzig. On 30th June and 1st July, the project partners gathered to discuss the progress in the work packages and the status of the demonstrators. The project partners are developing tools to help Web users, companies and organisations find and exploit geospatial data. The primary use cases are tourism e-commerce and supply chain management, but the tools can be applied in many more scenarios.
The GeoKnow team meets in Belgrade for the plenary meeting. During the 2 days the team discusses the achievements since the 1 year review meeting. Besides the ongoing improvement of the various tools the team discusses the topic of benchmarking and quality assessment. A key focus of benchmarking is on Virtuoso store, Facete and Mappify, LIMES, FAGI, GeoLift and TripleGeo. Results of the benchmarks will be published on https://github.com/GeoKnow/GeoBenchLab.
On the second day and the break-out sessions each individual work-package was thoroughly discussed and next steps were defined. Some of the findings are:
- Dashboard requirements and batch processing
- Parallelisation of LIMES process
- Notification and subscription service
- Mobile version for smart phones and tablets
- More free datasets that can be used for the use cases
The GeoKnow project just went through the Year 1 Review and straight after that meeting the consortium met for the plenary meeting on the 29 January 2014 in order to plan the activities within the work packages for the 2nd year. Below you will find the major high level information about the tasks that we will tackle in the upcoming months.
Work Package 1: One of the tasks will be to work on an authentication system that will connect the workbench with the underlying components. The initial version will be graph-based and tightly connected to the Virtuoso Graphs Security function. A first version should be ready by end of May 2014. In an extended version we will add multiple login options such as G+, Web ID etc. Another important issue is to work on scalability and therefore work on the benchmark such as the slippy benchmark, the BSBM BI load and harmonise this with the LDBC. Within the GeoKnow Workbench we will develop a dashboard and simple Workflow system that will allow the specification of processes that are interlinked and will be executed in the background and provide the monitoring of those processes.
Work Package 2: The team will work on improving algorithms related to geospatial dimensions that will support queries in relation to the different use cases. Most of the improvement will be within Task 2.6. InfAI and Athena will elaborate on on enhancements for the TripleGeo module that will include topics such as schema mappings and better support for the transformation process. Those activities are mostly a part of Task 2.7 and will also help to improve the scalability of the GeoKnow system.
Work Package 3: A high workload is related to task 3.2 and the integration of the tools such as FAGI, Facete and LIMES inside the GeoKnow Workbench. The integration work consist two parts. The first mainly concerns a tight technology integration and the second takes care of user interactions. Within task 3.3 major work will be devoted to the improvement of the GeoLift module and adding new datasets into the spatial mapping process. This will help to enrich the RDF datasets and create a larger knowledge base. Precision and and comparison of geospatial features such as metrics for polygons are the core challenges within task 3.4 that have to be solved and implemented. Based on those new implementations the link discovery process within LIMES has to be adapted and evaluated.
Work Package 4: Within WP4 the focus is on the visualisation and interaction with the users. IMP will add components that support mobile devices and a major step is to fully integrate the Facete and Mappify module inside the workbench. An improvement in filtering and selection will help users to navigate faster to the content. Further we have to solve the integration challenge of CubeViz and Facete. The above modules have to satisfy the requirements for knowledge engineer users but also for end users that will use just simplified UI versions to navigate and explore geospatial data content in the browser or on a mobile device.
Work Package 5: The main objective is to present at the next plenary meeting a supply chain scenario that is already using large portions of the GeoKnow workbench and have a real use case feedback.
Work Package 6: The real use case from the travel industry will do a full benchmark including the usage of the workbench on the transformation process including external and internal datasets. Further within this task will be a test where visualisation components will be deployed to the end user portals. Such a benchmark will provide feedback in scalability and handling and therefore provide deep insight about functional requirement, architectural issues and performance.
Work Package 7: Part of the dissemination work is to launch the GeoKnow workbench at two major events in March 2014 and attract a community that will start to use and test the system. Based on this we will be able to collect additional information that will help to tune the system. We will also provide documentation and tutorials on the Web page and continue to closely work with other funded projects.
Work Package 8: InfAI will continue the excellent work on managing the project, collecting and submitting the reports on time and hosting an advisory board telco. The next plenary meeting is set to take place in June 2014 and will most likely be organised by IMP in Belgrade.
Summary and Outlook
There are many small challenges that have to be solved but the major objective is to improve the overall integration of tools, run benchmarks with real use cases and work on scalability.
In the last days of July, the second meeting of GeoKnow project took place in Athens. GeoKnow members had the opportunity to meet again after the Leipzig kick-off meeting, discuss the work performed during the first 7 months of the project, as well as fix the next steps. Apart from that, our fellow partners had the chance to strall around some of the most historic and picturesque sites of Athens, like the old and the new parliament buildings, the National University of Athens, the Temple of Olympian Zeus, the Monastiraki and Plaka quarters and Acropolis, and taste some of the most iconic Greek dishes!
The first part of the meeting focused on performed work. Advances from each Work Package were presented, feeding discussions about (a) integrating currently developed tools, (b) utilizing these tools to manage/process use case datasets and (c) resolving research issues that had come up and enhancing the functionality and efficiency of the developed solutions. All partners agreed that the project advanced significantly since December, as the first tools for managing and processing geospatial, RDF data have been already developed, and very informative reports about the state of the art on geospatial and RDF data management, benchmarking and system requirements have been published as well.
Several discussions, during both meeting days, revolved around the system architecture and, specifically, the GeoKnow Generator (GKG). Concrete decisions were made about the GKG backend that will include a set of loosely integrated components for consuming, processing and exposing geospatial, Linked Data, based on Virtuoso RDF store. We also considered issues regarding user management, GKG’s front end, workflow processing and implementation of gathered system and user requirements.
With respect to geospatial information management, where GeoKnow has already provided solutions for transforming and exposing conventional geospatial data into RDF data (Sparqlify, LinkedGeoData, TripleGeo), all partners agreed that there is the potential to build (based on the work performed in Tasks 2.1 and 1.3) a timely geospatial RDF benchmark, that will be able to test efficiency and functionality capabilities of today’s RDF stores with geospatial support. Also, next steps were discussed, with emphasis on further optimizing the geospatial query capabilities of the underlying RDF store (Virtuoso).
As far as semantic integration of geospatial data is concerned, tools developed within Geoknow for enriching and interlinking geospatial RDF data (GeoLift, LIMES), were, at first, presented to the consortium. These tools triggered further discussions about the fusion and aggregation solutions currently under development and design, as well as how these tools can directly be tested and utilized into processing commercial datasets from the use case partners. Finally, a large part of our discussions was dedicated to quality measures and quality assessment of geospatial data; although these tasks are due to later periods in the project, they are of high importance for all the functionality being built in Work Package 3, since quality indicators of datasets can constitute valuable input for processing such as interlinking and fusion.
After the presentation of (implemented or under development) GeoKnow tools for visualization and authoring of geospatial RDF data, such as Facete, creative ideas were exchanged, discussing both detailed technical implementation solutions and desired functionality for end users. Some important aspects that were considered are the implementation and functionality of spatial authoring, the issue of public and spatial Linked Data co-evolution and the potential for spatial-social networking. Again, the discussions considered the use case scenarios of the project, that is, how the offered functionality can serve commercial and industrial Linked Data management and visualization needs.
In conclusion, during the GeoKnow meeting in Athens all partners exchanged interesting ideas about ongoing and future work and set more concrete objectives to achieve through the next months. We thank all the GeoKnow members for attending and contributing to this constructive meeting!
Last week (January 16th and 17th), we were warmly received by the InfAI team at one of the most interesting architectural buildings in Leipzig – the Paulinum at Universität Leipzig – to kick off the GeoKnow project. The meeting brought several discussions around the GeoKnow objectives to the table. Besides general discussions, in which everyone presented themselves and his/her role within the consortium, parallel meetings were organised where we discussed about research on spatial data as well as user scenarios and requirements within GeoKnow.
Geospatial Data Research
Research work package meetings were very interesting with discussions revolving around:
- How to benchmark geospatial data: Geo-spatial data storage is a challenge considering that there are points, lines and polygons describing data along with the usual structured information, which end up being stored in very large databases. Large datasets are combined with powerful distributed querying, which will be benchmarked in the project.
- Data fusing and link discovering: There is a lot of geospatial data and non-geospatial data that we can use on our scenarios. Therefore, we investigate efficient tools for fusing data and link discovering to enrich both geographical and not geographical data.
- Visualisation paradigms: visualisation requires to be as responsive and lightweight as possible – we discussed how to optimise data and data storage to cope with these objectives.
GeoKnow Generator and User Scenarios
Also during our kick-off, we discussed about the GeoKnow Generator. This framework is aimed to be used by anyone who wants to generate geospatial semantic web applications rapidly, and integrates the components (for querying, linking, browsing, editing, quality assurance) we are developing in the project. We pictured the first architecture of the GeoKnow Generator and the first user requirements according to our use case scenarios:
- Supporting supply chains: manufacturers need to improve information flow of material shipping from suppliers to its destination. Since they usually work in a just in time strategy, they need to avoid and anticipate any possible delay in the arrival of their materials. In our meeting, we could identify relevant requirements such as to be able to visualize real time data (route tracking material) and static data (supplier data) together with semantically enhanced message systems.
- Connecting E-Commerce information systems: Internet-based travel portals handle data of various types and sources (e.g. CRM system, travel agency products, POIs, hotel data, flight connections, etc) where linked data can overcome the problem of scalability and data integration. Improved management of geographical data in this use case provides benefits for travel agencies, customers and service providers since the geographical aspects will help in the discovery and matching of products and customers.
The meetings around use cases helped us to start a user requirements definition. However, we are also interested in hearing other use cases exploiting linked spacial data. You can drop us a message at firstname.lastname@example.org if you are interested in exploiting geospatial linked data in your use case.
We want to thank to all partners for joining and contributing to the meeting and look forward to interesting research and applications within the project. Come back soon to read more about Geoknow.