W3C Logo Open Geospatial Consortium Logo

Linking Geospatial Data

5th - 6th March 2014, Campus London, Shoreditch


Minutes: Day 1, Day 2

Participants are strongly encouraged to join the chat channel throughout both days. This will be how we take minutes from the event, which will be published as raw notes to back up the more considered report. You are welcome to make comments in the chat room, correct errors, add more detail to the chat log etc.

Tweets using the hashtag #lgd14 will be archived on this site.

A zipped archive of all the papers is available for easy download.

Doors open, quick coffee
Welcome & Introduction, Phil Archer (W3C) and Ed Parsons (Google)
Working With Standards Chair: John Goodwin; Scribe: Raphaël Troncy; (10 minutes per speaker plus 30 mins discussion)
Using INSPIRE data on the Web Clemens Portele Interactive Instruments [paper] [slides]

This paper builds upon the talk "How to use INSPIRE data" given at the INSPIRE Conference 2013 and describes experiences from the ongoing work in the European Location Framework (ELF) project. It discusses patterns how the usability of INSPIRE data on the web can be improved and identifies open challenges that we are facing. This talk will focus on two technical aspects. The first aspect relates to requirements and patterns for reference geography on the web that supports linking other data to that geography. The second aspect is about bridging the different platforms and technologies.

Developing Ontologies for Linked Geospatial Data, Kerry Taylor, CSIRO [paper] [slides]

We compare experiences in modelling geographic information in OWL by two very different methods. Both methods have a root in OGC standards, but one followed a semantic web development style and the other a UML-flavoured Model Driven Architecture style. We ask how much of the OGC UML modelling should be preserved in the transition from UML modelling to OWL modelling and from XML to open linked data.

Enriching the German National Library’s Linked Data Service with Geographic Coordinates : Approach and Lessons Learned (so far) Lars G. Svensson, Deutsche Nationalbibliothek [paper] [slides]

The German National Library (Deutsche Nationalbibliothek, DNB) publishes geographic coordinates for approx. 40,000 geographic entities and 20,000 maps and charts. The data is available in several formats, including MARC 21, CSV and RDF. This paper discusses the design issues considered when choosing which RDF vocabulary to use and why the preliminary choice is to use geosparql and simple features. The paper is concluded with some examples of the current implementation and an outlook at some future work.

Lightning Talks: Reference Data Chair: Martin Tuchyňa, Slovak Environmental Agency [paper] ; Scribe: Hadley Beeman
(5 mins + 5 mins Q&A each)

Paper is addressing potential of spatial related items registration via Registries and registers with focus on environmental domain Feature catalogue register deployment. Main motivations for overall registers use as well as description of specific case from Slovakia indicated the needs and specific requirements for deployment of this type of register with solutions currently available on the scene. Basic description of two identified registry platforms is provided with focus on potential Feature catalogue registry implementation is summarised with outlook for further activities foreseen to be undertaken.

Interoperable Registers and Registries in the EU: Perspectives from INSPIRE Andrea Perego/Michael Lutz, EC/JRC [paper] [slides]

INSPIRE is a EU-wide data and service infrastructure for the cross-border sharing of environmental data and for their use in support to policy making. This paper introduces the context, requirements and issues for registers and registries in INSPIRE, including persistent identifiers, versioning, multi-linguality, extensibility, linking and alignment with existing registers and cross-sector interoperability and re-use. In our presentation, besides highlighting open issues relevant not only in the scope of INSPIRE, we will report the results of an INSPIRE workshop on registers/registries taking place on 22-23 January 2014.

RDF and PIDs for INSPIRE: a missing item in ARE3NA Diederik Tirry, Katholieke Universiteit Leuven [paper] [slides]

Our proposed presentation will outline intermediate results of a study in the context of geospatial data sharing across borders and at European level. The study is aiming to develop a common approach to generating common RDF schemas for representing INSPIRE data and metadata, as well as guidelines for the governance of persistent identifiers (PIDs). These are important elements for enabling the re-use of INSPIRE data in other sectors, in particular in e-government. The results of the study may feed into a proposal for additional encoding rules and guidelines for INSPIRE and will be performed in close collaboration with the INSPIRE Maintenance and Implementation Group and the ISA Programme’s Spatial Information and Services Working Group.

Research Data Alliance: Research Data Sharing without barriers Herman Stehouwer, Max Planck Institute for Psycholinguistics [slides]

Several areas that are being addressed in the RDA are relevant to the linked geospatial data, and linked data communities.

One development that is beneficial to linked data is that of persistent identifiers. Within RDA the PID information types group is working on the problem of standardizing the information associated with PIDs and defining a common API to access that information regardless of the underlying PID system.

Take, for instance, the group on citation of dynamic data. This group works on the referability of data that is mutable. Some common reasons for the mutability of data are: temporary loss of networking with sensors, re-calibration of sensors resulting in a correction of past measurements, unreliable dataflows in combination with real-time decision making, etc. It is essential to get exactly the same data again at a later date. This specific group tries to develop a system for precisely such circumstances. An other opportunity offered by the standardization effort in the dynamic data referability is to foster the use of the temporal data associated to the geospatial ones. Because a common form of dynamic data is time series data in which measurements are obtained on a regular schedule, with a well-defined sample rate. These are quite common in geosciences, from seismic maps and weather forecasts, to gps-based traffic information systems. The usage of dynamic geographic information is also the base for many modern applications providing environment monitoring and alert systems, such as pollution sensor network or tsunami satellite and marine monitoring. The RDA activities to set a common ground across these different scenarios can pave the way for smarter Web applications and enforce the Linked Data approach. For example, it would be easy to associate RDF triplets to PIDs attached to geospatial data, once defined a common vocabulary and a common set of information types.

It is clear that the linked data community can benefit from the knowledge gained and data exchange that is enabled through these groups. Vice versa, the RDA can benefit greatly from active linked data involvement to help steer the outcomes to more generally applicable.

Oceans of Linked Data? Adam Leadbetter, British Oceanographic Data Centre [paper] [slides]

Work has been undertaken to publish Linked Ocean Data, using the controlled vocabulary content of the NERC Vocabulary Server as the central, common resource. However, it is recognised that marine science data are geospatial by their very nature and that this work remains largely explored.

This position paper details the Linked Ocean Data work, and examines the possibilities for Linking Geopatial Data in the marine science domain.

Smile For The Camera (Earth Observation) Chair: Alex Coley; Scribe: Stuart Williams; (3 x 5 minute presentations as intro to panel theme)
The MELODIES project: Exploiting Linked Open Geospatial Data Jon Blower, Reading University [paper] [slides]

The MELODIES project (Maximising the Exploitation of Linked Open Data In Enterprise and Science) is developing eight innovative and sustainable services, based upon Linked Open Data, for users in research, government, industry and the general public in a broad range of societal and environmental benefit areas. The project is in its early stages and we therefore offer a presentation that describes our aims and plans, with the intention of stimulating discussion. The project presents a number of key challenges to the use of Linked Data, including integrating diverse data sources, interoperability with spatial data infrastructures, handling time-varying data, handling raster data and more. We would welcome the opportunity to discuss these challenges in this forum.

Expressing weather observations as Linked Data; ISO 19100 geographic information meets semantic web head on Jeremy Tandy, Met Office [slides]

As a member of WMO (on behalf of the UK), and participant in the World Weather Watch, the Met Office actively supports the international standards efforts within the meteorological community. The Met Office has a strong desire to implement standards such METCE and is committed to meet the legislative requirements from INSPIRE. At the same time, the Met Office contributes to UK Government open data initiatives which, ultimately, includes the aspiration to provide content as Linked Data compliant with “semantic web” standards from the World Wide Web Consortium (W3C).

Work is underway within ISO/TC 211 to bridge the gap between geographic information standards and the semantic web. Draft international standard ISO/DIS 19150-2 “Geographic information - Ontology - Part 2: Rules for developing ontologies in the Web Ontology Language (OWL)” provides a set of rules for mapping Abstract Specifications defined within the ISO 19100-series of geographic information standards, such as O&M2, to OWL.

Working versions of OWL ontologies derived from these Abstract Specifications using the (draft) rules are hosted by CSIRO at http://def.seegrid.csiro.au/isotc211. A snippet of weather data for Exeter International Airport (extracted from the Met Office’s “Weather Observations Website”) encoded in RDF using these OWL ontologies provides an illustrative example.

In addition to O&M2, the Abstract Specifications cover topics such as spatial schema (ISO 19107:2003), temporal schema (ISO 19108:2002), metadata (ISO 19115:2003 & ISO 19119:2005). Many of the Classes and Properties expressed in the OWL ontologies overlap entirely, or partially, with RDF/OWL ontologies in common usage within the semantic web - many of which originate from W3C:

  • O&M2 may be partially mapped to the Semantic Sensor Network Ontology; as described in "An explicit OWL representation of ISO/OGC Observations and Measurements” (Cox, 2013, 6th International Workshop on Semantic Sensor Networks)
  • spatial schema may be partially mapped to the Simple Feature Geometries and GeoSPARQL
  • temporal schema may be partially mapped to “time” and “timezone” ontologies
  • metadata may be mapped to Dublin Core, with “Responsible Party” relating to FOAF.

The example weather data snippet once re-encoded using OWL ontologies in common usage illustrates the difference in approach.

To further align the geographic information and semantic web communities, it is important that existing OWL ontologies in common usage are recognised, and that the OWL ontologies derived from the geographic information Abstract Specifications incorporate existing works where deemed appropriate. Assessment of potential alignment between geographic information Abstract Specifications and existing OWL ontologies will be of interest to all communities who have developed Application Schema compliant with ISO 19109 (General Feature Model) to describe their information resources and are beginning to evaluate the benefits of Linked Data publication.

Given the effort required, the benefit from broad-based peer review and the likely cross-community interest, community-driven work item may be the most appropriate way to resource such an activity.

Data Discovery: Unleashing the Power of Geospatial Linked Data Dicky Allison, Woods Hole Oceanographic Institution [paper] [slides]

BCO-DMO serves a wide variety of oceanographic related data, including but not limited to biological, chemical and physical oceanography measurements. We routinely deal with CTD data, biological abundance, meteorological data, nutrients, pH, carbonate chemistry, PAR, sea surface temperature, heat and momentum flux, sediment composition, trace metals, primary production, and pigment concentration measurements, and with images and movies. In addition to a text-based interface, we offer an OGC-compliant geospatial interface (MapServer) to enable the user to find the data, evaluate its fitness-for-purpose, and download and re-use the data for his or her own analyses. BCO-DMO manages data that are heterogeneous in the extreme, and therefore one method of geospatial display does not work for all. Because the data are not always the same kind, format, scale, or magnitude, each ‘quick look’ at the data using the map must be individually tailored. Describing each dataset for the mapper is done via specific mapping parameters assigned by the data managers and added to the metadata associated with each dataset and stored in a MySQL database. It is a challenge to communicate with the user how to make the displays that we hope will add value to the data and we address this challenge by providing aids in the form of descriptions, tips, and help boxes in as many places as possible. We are always asking: Are we giving the user what they really want? We believe people want to be able to look at data from multiple sources, but interoperating with other sites is difficult to achieve. Can we link to other data and use our own display methodology? And even if we could, would it be at the same scale or precision or quality? There are many reasons for linking geospatial data from multiple sources but there are cautions as well.

Chris Little, Met Office
  1. Meteorology has been exchanging data, products and information globally for well over a century, and has a strong global community of committed professionals, and informed amateurs too, as well as most of the globe’s population as users of some of the information.
  2. Global standards have existed, and been maintained and updated over decades within the community. These standards and Best Practices cover data formats, processes, training, maintenance procedures, presentation, semantics, calibration, etc.
  3. National and regional variants of all of above, and non-standard, proprietary versions, have existed, but these have largely disappeared in the face of true globalisation.
  4. Most of the data and products are geo-spatial, but coordinate systems may not be considered conventional, such as vertical coordinates of geo-potential height, hectoPascals (think millibars) or even potential temperature, which is not even single valued. Assuming that the earth is spherical, rather than a geoid or oblate spheroid is nearly always good enough.
  5. Data standards include:
    1. Telegraphic orientated textual formats, not really suited to machine processing, but were very compact and semantically dense for transmission over limited bandwidth. The semantics are defined and maintained in manuals of code tables (basically data dictionaries) in the 6 working and official languages (Arabic, Chinese, English, French, Russian and Spanish)
    2. Table driven formats that were flexible and maintainable from the software point of view and form the aspect of semantic content. These formats encompassed point (observational) data, trajectories or soundings, and gridded data.
    3. Presentation styles for maps and other graphics were well defined and are pervasive and understood among maritime, aviation and professional meteorological communities. Map Projections are typically conformal (Equal-Angle) rather than Equal-Area, to address the needs of navigation and polar views.
  6. Attempts to persuade wider, extra-community uses of these standards over the last 20 years have, by and large, failed, as the detail and complexity needed to address ICT constraints, safety critical issues, sustainability and legacy compatibility have proved too great a barrier to other users.
  7. More than ten years ago, the global governing bodies of Meteorology, Hydrology and Oceanography recognised the problem and embarked on a long term initiative to make their data and products much more accessible and available, by Implementing the WMO Information System (WIS), and the WMO Integrated Global Observing System (WIGOS). These efforts also feed into the intergovernmental body GEO and its Global Earth Observation System of Systems (GEOSS).
  8. Current activities are:
    1. A comprehensive, distributed global catalogue of all meteorological data, products and services, based on ISO 19115 DAR metadata records, OAI-PMH harvesting and ISO23950&SRU1.3 search. Various technologies are being used but not mandated.
    2. Upgrading the global meteorological safety critical extranet to cope with the expected extra traffic.
    3. Develop underpinning formal conceptual models to enable largely automatic conversion between data formats and future data serialisations.
    4. Establish authoritative and sustainably maintained registers of semantic content.
    5. Influence OGC and other global standards bodies to ensure the community needs are met and increase the availability of our products and services.
  9. What are we doing in detail:
    1. OGC WMS1.3 Best Practice for Time and Elevation in Meteorology and Oceanography. Done.
    2. Made weather and aviation symbols available in SVG and GitHub, rather than paper manual. Done.
    3. OGC WCS2.0 allows slicing and dicing of 4D data, if data gridded and homogeneous. I.e. It is completely regular in each dimension. Done.
    4. Proposing WCS2.0 Extension for Meteorological data that may not be completely regular and homogeneous, using bit maps in metadata to specify missing ‘layers’.
    5. Just established OGC Temporal Domain Working Group to establish Best Practice for time (2000, 2038, 1582/1923, space weather, GPS and Leap seconds, etc) and formally register temporal CRSs.
    6. Trying to establish WCS Tile Service, to serve local gridded and other ‘containers’ to support mobile users and devices. Meteorological data is usually produce in large batches, not ‘streamed’. Tiles maybe like Map Tiles (WMTS) or in time or the vertical, or some other ‘abstract’ dimension.
  10. Issues:
    1. Volume: a terabyte per day and growing, from each of the big global centres;
    2. Volatility: out of date every day;
    3. Authoritative voice and managed data (“This morning’s flood warning has been superseded”);
    4. Long haul, which technologies and institutions to pick?
    5. Leading/bleeding edge versus trailing edge. We do need Burkina Faso’s data, every hour, never mind South Sudan, Timor Leste, DRCongo, etc. Global development and technological neo-colonialism;
    6. Both remote and in situ data needed. It cannot all be done from space or drones;
    7. 2D versus 4+D. Existing geospatial interoperability standards are largely still 2D, static and flat earth.
  11. Problems:
    1. Catalogue or linking points to a single big blob with some internal structure E.g. point cloud, or 3 or 4D grid, regular or slightly sparse. How can we get inside the blob for further meaningful discovery/access/retrieval? Akin to the Google problem: here are the first 20 results of 10m hits. Preferably without the user having to learn domain specific terminology or concepts.
    2. Can we use the formal conceptual models and semantic registries to constrain the search into the blob?
    3. Can we distinguish between the formal conceptual model structures and the actual implementation structures, which may be different, especially when accessing archived or legacy material?

I am happy to expand further if some of this does not make sense to an outsider/other insider.

Massimo Zotti, Planetek Italia [paper]

The availability of remotely sensed data acquired by satellite, airborne or UAV platforms, increases continuously, with a parallel increase in the availability of open satellite datasets. More and more in fact the space agencies data policies go in the direction of providing open access to the data collected by satellite platforms for Earth observation (EO). Those information sources bring to the imagery big data challenge: a growing availability of data must correspond to a greater ability to automatically and quickly extract and share information, in the most effective way.

Transforming the value-added information extracted from satellite images in Linked Open Data (LOD) responds to this need. The analysis may be fostered by crossing data from different sources of information, accessible via the Internet, which refer to the same area of interest.

The use of standardized and common data structures and semantics allows to define common ontologies: this way the information may express its full value, without the intervention of human operators. INSPIRE, the European Directive which defines the specifications for 34 data themes (hydrography, administrative units, parcels, buildings, etc.) which all Member States have to comply, resolves the problem of the heterogeneity of formats, data structures and semantics.

The exploitation of value added information, quickly extracted from Earth observation data, and compliant with the specifications defined by the European regulations, can be by far enhanced thanks to LOD, thus overcoming the inefficiencies and delays that still today restrain the real value of such an extraordinary sources of information.

This article aims to identify some business opportunities for Earth observation companies providing the so-called downstream services within the European program Copernicus, related to LOD and arising from the imminent availability of the Sentinel satellite data.

Bente Lilja Bye, BLB [paper]

Geospatial information serves as an instrument in combining various kinds of data, both topical and temporal, gluing together data collected by different technologies be it ground measurement or from space.

Further development of techniques that will be able to combine and make use of both authoritative and non-authorative data such as data collected from social media will be helpful in a number of applications areas where knowledge about reliability, quality and source is of importance.

Will open solutions provide the necessary trust in the combinations, and do we have data policies that will allow access to geospatial information required for establishing trust in real-time/near real-time information (such as event related data from social media)?

BLB works on bridging the institutional gap between end-users (private and public) and science with focus on Earth observations. BLB has experience from quality assessment of near-real time data, data policy development, certification and labeling of data and various geodetic data in early warning systems applications.

Tony Bush, Ricardo-AEA [paper]

The quality of the air we breathe is a health concern for many. It can be a major factor in how much we enjoy life - whether we go for a run, ride a bike, walk the dog or indeed whether we go out at all. Since 2010 the UK has been leading work to make air quality data open, interoperable and available – to better inform, protect and serve the public interest. Uniform Resource Identifiers (URIs) for air quality data are fundamental to achieving this and bridging the gap between INSPIRE and the semantic web.

In 2008 EU regulations for location-based data (INSPIRE) were formally adopted within EU air quality regulations. Since then the Ricardo-AEA and UK Government has been leading the way in a move to open data for air quality. Media channels such as telephone lines, HTML pages, Twitter feeds and RSS have been the de facto requirement for air quality information-sharing for some time now. The UK is the first to replicate these channels using INSPIRE-based, open-standards.

In its infancy the INSPIRE initiative was all about location and geography, but to realise its full benefits the focus must also be on data. Data to inform, empower and serve policy at all levels.

The past 3-years has seen a significant change in the way air quality data are managed and published across Europe, pushing the boundaries and time-scales of standards development for data publishing. UK has been at the forefront of this work. This paper presents the challenges and benefits of being the first INSPIRE Annex III data theme creating a domain specific vocabulary, data model, application schema and to implement http URI patterns that will support the UK Government’s journey towards web services for open air quality data and placing of “air quality” on the semantic web.

Lightning Talks: Project News Chair: Bart de Lathouwer; Scribe: Phil Archer; (5 mins + 5 mins Q&A each)
The CLEAR Project, Christopher Baldock, Trucost [paper] [slides]

The CLEAR Info team would like to present aspects of the EU funded project which aims to link and integrate global company-level environmental data to give investors, regulators and the companies themselves insight into environmental and financial risk. The CLEAR Info project is led by the English Environment Agency and has a number of project partners that include Trucost, University of Southampton, Eurosif as well as a number of other organisations.

One of the main aims for the project is to be able to aggregate various environmental datasets to the parent companies that are ultimately responsible for its performance. The Environment Agency has conducted trials of its integrated data platform, which contains data collected by the Agency, with companies such as Veolia and GlaxoSmithKline and we are now further developing our systems take into account external environmental datasets (GIS data and so on) to meet the needs of the stakeholders mentioned above.

The presentation that we would like to give will demonstrate the Environment Agency's data integration platform and introduce the processes, barriers and challenges regarding the integration of external company-level and geo-spatial datasets.

Open Data for Real Estate Business, Stepan Kafka, Czech Center for Science and Society [paper] [slides]

The contribution presents two solutions based on Open Data which are targeted mainly at real estate businesses. The research and developments were performed mainly within the EU FP7 plan4business project – a Service Platform for Aggregation, Processing and Analysing of Urban and Regional Planning Data. The platform enables harmonisation of data into a seamless, homogenous, constantly growing and updated trans-border dataset. The platform enables spatial analyses across European datasets. The platform serves not only as a catalogue of planning data but also as their integrator enabling users to search, view, analyse and download spatial planning data on European and regional levels. The main objectives are the automation of harmonisation processes and possibilities of complex analyses.

One of the major plan4business outcomes is the Thematic Map Viewer. This app visualises data stored in the database in a user friendly way. Due to the fact that the database contains many data-layers, a grouping of these layers took place - into thematic compositions. By now, we have created about 50 compositions. Most of them are related to socio-economic and demographic indicators such as GDP, average monthly salary, unemployment rate, employment structure (by sectors), local human development index (LHDI), population size and density, net migration and natural growth and age dependency ratio.

Another outcome is the Location Evaluator app. The main concept of this app is to have a possibility to prepare reports in the form of a general template. These templates are then inserted into the database and finally they are connected to the user interface. One of the main objectives of this app was to make the preparation of such templates easy also for non-programmers.

The GeoKnow Generator: Managing Geospatial Data in the Linked Data Web, Claus Stadler, University of Leipzig [paper] [slides]

Within the GeoKnow project, various tools are developed and integrated which aim to simplify managing geospatial Linked Data on the web. In this article, we give a first presentation of the GeoKnow Generator, which is the platform providing a light-weight integration of those tools.

Lunch Kindly provided by our hosts, Google
Linked Geospatial Data Chair: Stuart Harrison; Scribe: Jeremy Tandy; (10 minutes per speaker plus discussion)
Linked Location Data as a Service, Stijn Goedertier, PwC [paper] [slides]

Geographic information is provided in heterogeneous formats, creating technical and semantic barriers that hinder data consumers to combine data from various sources. Linked Data design principles can help alleviate these barriers and realise three generic use cases for data consumers to consume location data as linked data. We have demonstrated this in a Linked Data pilot that integrates address data from five different public sector organisations in Belgium. The pilot demonstrates that a Linked Data layer can be built on top of an existing geospatial implementation with a minimum of effort. It also shows that URI sets for INSPIRE spatial objects and spatial things can accommodate both XML (GML) and RDF representations.

Linked Earth Observation Data:The Projects TELEIOS and LEO, Kostis Kyzirakos, CWI [paper] [slides]

Lots of Earth Observation data has become available at no charge in Europe and the US recently and there is a strong push for more open EO data. Open EO data that are currently made available by space agencies are not following the linked data paradigm. Therefore, from the perspective of a user, the EO data and other kinds of geospatial data necessary to satisfy his or her information need can only be found in different data silos, where each silo may contain only part of the needed data. Opening up these silos by publishing their contents as RDF and interlinking them with semantic connections will allow the development of data analytics applications with great environmental and financial value. In this paper, we present the advances in the areas of Semantic Web and Linked Data achieved by the European project TELEIOS and give a short overview of the goal of the new European project LEO.

Linked Data and geoinformatics, Frans Knibbe, Geodan [paper] [slides]

Recent developments in the domains of Linked Data (or the Semantic Web) and Geoinformatics have been largely independent, but mutual interest is clearly growing. Both worlds have a lot to offer to each other, and can make each other stronger. But there are a few gaps between the two that have yet to be bridged. In this document some of those gaps are described, together with possible ways of narrowing or even bridging the gaps.

Ecuadorian Geospatial Linked Data, Boris Villazón-Terrazas, iSOCO [paper] [slides]

Much of the attention lately has been on geospatial data and their combination with Linked Data. In this paper we present our approach for creating the Ecuadorian Geospatial Linked Data repository, pointing out some existing efforts, and open issues we have to cope with in order to make it easier the publication and exploitation of Geospatial Linked Data.

Lightning Talks: An Alternative Projection Chair: Leigh Dodds; Scribe: John Goodwin; (5 mins + 5 mins Q&A each, plus discussion time)
Geospatial Data Integration with Linked Data and Provenance Tracking, Andreas Harth, Karlsruhe Institute of Technology [paper] [slides]

We report on our experiences with integrating geospatial datasets using Linked Data technologies. We describe NeoGeo, an integration vocabulary, and an integration scenario involving two geospatial datasets: the GADM database of Global Administrative Areas and NUTS, the Nomenclature of Territorial Units for Statistics. We identify the need for provenance to be able to correctly intepret query results over the integrated dataset.

Geospatial Data and Linked Data, Mano Marks, Google [paper] [slides]
Modeling Geometry and Reference Systems on the Web of Data, Raphaël Troncy, Eurecom [paper] [slides]

For many years now, the web of data has been dominated with the use of only one Coordinate System (CRS), namely WGS84, to represent the localization of geographic objects on Earth. Reasons for its adoption is the simplicity of the vocabulary (few core classes and properties) and the fact that the vocabulary is described in a W3C namespace. Nowadays, with the Open Data movement, more and more publishers including governments and local authorities are releasing legacy data that is often geolocalized in a different coordinate system. For example, IGN in France in releasing data that is geolocalized using Lambert93, a Lambert conformal conic projection (LCC) when objects are localized on the France metropolitan area. In this paper, we propose two vocabularies that take into account geometries defined in different coordinate systems. We provide as well mappings with existing vocabularies to ensure compatibilities with existing implementations. Finally, we provide a REST service that supports the conversion of coordinates between several CRS.

It's About Time Chair: Hadley Beeman; Scribe: Andreas Harth; (10 minutes each plus discussion)
Modelling and managing temporal changes to geographic datasets, Steve Peters, Department for Communities and Local Government, Bill Roberts, Swirrl [paper] [slides]

This paper outlines the common problem of handling temporal change to geographic LinkedData datasets, with thoughts on how it might be resolved.

We would welcome feedback, and would like to work with others to agree a best practice or standard that could be applied consistently by data owners with similar data.

Towards a geotemporal semantic infrastructure for Dutch cultural historical data, Rein van 'T Veer, Digitaal Erfgoed Nederland/Erfgoed en Locatie [paper] [slides]

Dutch Cultural Heritage is building on the Heritage & Location project (Erfgoed & Locatie) to bring its content to the semantic web. Heritage & Location is looking for strategies to make this happen, not only to make cross-domain searches, but also to spatially and temporally enable the cultural heritage data. The current state of affairs, however, leaves the project in a situation where the desired search capabilities using combined semantic, temporal and spatial functionality has no clear-cut server stack candidate. In theory, the GeoSPARQL functions provide all the necessary spatial functions, but current implementations seem to lack either stability, scalability, documentation, semantic reasoning or a combination of these. Perhaps a solution can be found in combining several components that specialise in any (combination) of these requirements, or maybe even a fully developed single package is waiting around the corner. We greatly welcome any suggestions for upcoming or existing solutions.

From the Trenches, Rob Warren, Big Data Institute, Dalhousie University [paper] [slides]

This paper reports on the experiences of building a linked geo data coordinates translation API and some of the issues that arose. Beyond the basic capacities of SPARQL, a specialized API was constructed to translate obsolete British Trench Map coordinates from the Great War into modern WGS84 reference systems. Concerns over current methods of recording geographic information along with accuracy and precision information are reported on. Open questions about managing the opportunistic enrichment of geographical instances are discussed as well as the scalability pitfalls therein.

Discussion including Susanne Rutishauser [paper]

Over the past few years, archaeology has experienced a rapid development in geophysical prospection and remote sensing techniques. At the same time, the focus of archaeological research has shifted to landscape development and human interaction. To impart the results, new methods and techniques are necessary. Virtual globes such as Google Earth offer fascinating methods of giving interested amateurs the possibility to interactively explore ancient cities and landscapes. Thanks to the increasing usage of GIS in cultural heritage, the implementation of interactive three dimensional learning opportunities becomes less and less tedious, but the non-linear narrative story telling medium demands for a special adaption of the content. This paper summarizes the experience gained during the realization of the “Virtual Cilicia Project” and outlines the future potential of virtual globes in the field of cultural heritage.

Closing Remarks, Phil Archer
End of Day 1. See side panel for details of networking event at Open Knowledge Foundation.
Guests include Matt Amos from Open Street Map Foundation

Doors open, quick coffee
A Different Set of Problems Chair: Ed Parsons; Scribe: Alex Coley; (10 minutes per speaker plus discussion, the 2 BGS papers share one slot)
Knowledge Graph of Places, Zhenning Shangguan, Pitney Bowes [paper] [slides]

Pitney Bowes, as a market leader in the GIS industry, possesses large amounts of geospatial data, covering business points, land parcels, road networks, indoor maps, Points of Interests (POI), 3D building structures, and geographical boundaries. Recently, data sources that generate geospatial data in the form of continuous streams, such as real-time traffic, social activities and GPS navigation trajectories, are also incorporated into our data universe. In theory, these datasets should enable us to provide rich contextual views of places in the physical world and enable new business opportunities (in the form of new data products and services), once properly combined and presented. However, reality is that even within the company itself, typical daily operations, such as discovering, exploring, merging, augmenting, and making good use of these geospatial data assets, quite often involve cumbersome steps and processes. This is the major motivation that drives the need for a linked geospatial data platform, which we call the Knowledge Graph of Places.

Linked Geospatial Data, and the BBC, Chris Henden, BBC [paper] [slides]

This paper aims to communicate the current state of play for geographic data and content for BBC Online. We want to show why place matters for the BBC audience, and provide a simple overview of the approach taken to date. Our hope for the conference is to provide any insights possible, find if we are in sympathy with others, and ideally to improve our approach.

Delivery and querying of 3D spatial models, John Laxton, BGS [paper] [slides]

BGS is the UK's main provider of geoscientific information, and the geological map was traditionally the main means of delivering this. Digital geological maps, linked to GIS spatial databases, contain a rich set of information and in order to deliver this using web services a data model and interchange specification (GeoSciML) was developed based on OGC standards. GeoSciML was used as the basis for developing the INSPIRE Geology data specification.

Vocabularies and ontologies available in SKOS RDF have been developed to constrain the GeoSciML fields.BGS has developed a linked data service for some map and ontology information and we are investigating further applications of linked data, such as cross-domain querying.

Geological maps are being replaced by 3D geological models and in the FP7 EarthServer project we are investigating the use of WCS and WCPS to deliver and query these over the web. Work in the project is looking at combining xQuery and WCPS.

BGS has also developed a range of web-based mobile applications and we are interested in extending the functionality of these through links to ontologies.

Experiences with implementing interoperable WFS 2.0 services using open source software – a UK INSPIRE Issues Paper, Tim Duffy, BGS [paper] [slides]

Issues with implementing interoperable INSPIRE complex property GML schema data using open source software trying to follow OGC standards are discussed.

Discussion including Peter Rushforth [paper] Tomaž Žagar [paper]

GeoGratis is the name of Natural Resources Canada’s web portal for open geospatial data and services. GeoGratis offers a diversity of data through multiple facilities. One of the important facilities is the GeoGratis API, which is an Atom Publishing Protocol service, offering a catalogue of ISO 19115 Metadata – North American Profile metadata records linking to hundreds of thousands of open data sets.

The GeoGratis API uses a hypermedia approach to delivery of linked geospatial data. The core formats are extended from the standard AtomPub hypermedia formats. Each resource is offered in multiple formats either via standard content negotiation or via typed links to API query resources which are programmed to return the advertised format. Supported formats include the ‘native’ XML-based formats of AtomPub, as well as JSON, KML, ISO 19115 XML, RSS and of course HTML.

StaGe is a web mapping application for merging and presentation of statistical and spatial data. It comprises StaGeSERVER, StaGeADMIN and StaGeVIEW.

The map is rendered in an HTML5 canvas according to the WMS image pixel value and the corresponding attribute value. This approach was chosen to unify the presentation of non homogenous data sets, e.g. some layers have 2 polygons with several thousand of points while other layers are grids (having several thousand – even 100 000 or more – quadrilaterals) and there is a demand to show several thousand of objects when the large map areas are being viewed. Another reason for the chosen approach was to move the majority of map processing from the server to the client, however, there is a question how this approach can be used in connection with standard OGC interfaces or Inspire Network services efficiently.

Putting The Data To Work Chair: Bill Roberts; Scribe: Stuart Harrison;
(10 minutes per speaker plus discussion)
RAGLD: The Rapid Assembly of Geospatial Linked Data applications, John Goodwin, Ordnance Survey [paper] [slides]

With more and more linked data and open data emerging there is a rising demand for a suite of application developers’ tools to make it easier to bring together, use and fully exploit this diverse data.

RAGLD, a collaborative project between Ordnance Survey, Seme4 and the University of Southampton, produced a set of tools, components and services making it easier to develop applications by helping to speed up and enhance the use of linked data and opendata.

Map4RDF-iOS: a tool for exploring Linked Geospatial Data, Alejandro Llaves, Ontology Engineering Group, Universidad Politécnica de Madrid [paper] [slides]

In this paper we describe Map4RDF-iOS, a tool that allows visualizing and navigating through RDF-based geographic datasets available via a SPARQL endpoint, as well as connecting that data with statistical data represented with the W3C DataCube vocabulary or sensor data represented with the W3C Semantic Sensor Network ontology.

The Android App Geohazard – Experiences with shared information on natural hazards; State of the Art and forward thinking, Martin Hammitzsch, GFZ German Research Centre for Geosciences [paper] [slides]

Geospatial information on natural disasters is publicly made available worldwide by various organisations via Internet platforms. On one side, these are authorities who deal directly with the respective natural event to prevent catastrophes and protect the population. On the other side, these are often scientific institutions that compile and provide information through research, and act as advisors to the authorities concerned. In addition, it has turned out that effective planning of relief and rescue measures requires both information provided by these authorities and feedback of the general public. Every citizen who has been directly affected by the event on-site can give valuable information as a potential witness. This is especially useful during the events, where witnesses can provide specific statements on the current impact on man and the environment.

During the last years, the crowdsourcing approach has gained the attention of users of modern communication and information systems. The term crowdsourcing describes the interactive, decentralized, and even ad-hoc collaboration of volunteers working on a common topic via crowdsourcing platforms. This approach is extended by mobile crowdsourcing and crowd-mapping which evolved in the quickly growing community of smartphone users and the ability of location awareness by devices and software tools. So smartphone users are enabled to compose and share reports immediately at the scene of natural hazards or even disasters.

The project Collaborative, Complex, and Critical Decision-Support in Evolving Crises (TRIDEC), co-funded by the European Commission in its Seventh Framework Programme, is focused on the management of crisis situations. Part of the project is the development of an application named Geohazard for the Android smartphone platform. The App is primarily designed to immediately sent eyewitness reports on natural hazards, to an instance of the crowd mapping platform Ushahidi, which is used repeatedly since the devastating 2010 earthquake in Haiti to collect eyewitness reports in natural disasters and thus optimally supplements the conventional sensors and sensor systems used for event monitoring. Crisis reports which include geographical information with event classification, description and photos can be directly visualized in a geographic information system on a crowd-map for the benefit of the crowdsourcing community and the individual user as well as national and international authorities.

Furthermore, this application enables access to continuously updated information for current natural hazards like earthquakes and tsunamis. The Geohazard App has been developed and implemented in such a way that different data sources, such as from BGS (British Geological Survey), USGS (United States Geological Survey), NOAA (National Oceanic and Atmospheric Administration), and GDCAS (Global Disaster Alert and Coordination System) can be incorporated into a central automatism. At present, data from more than 15 providers are used for the Geohazard App in order to give app users information on earthquakes, tsunamis, floods, hurricanes and volcanic activity. Currently, the combination and uniform presentation of data on various natural hazards from various sources in one GeoApp is unique and has pointed to several problems regarding interoperability.

An important issue is that a wide variety of data formats were offered by the different data providers. Even where same data formats have been used, i.e. RSS feeds, they have been used and extended differently, i.e. with additional data and with other formats carrying these data. Specifically a variety of metadata standards and non-standardized formats is used world-wide for these purposes without any common understanding or agreement. These circumstances were particularly challenging when developing the Android app. Several separate functions were implemented for data processing for each organisation to integrate their public data into the application. A uniform use of standards to shorten these processes and support interoperability is not only highly welcome but also a driving key factor while managing crisis events jointly with different players involved in a heterogeneous environment. This affects not only in terms of data representation but also in terms of semantics, e.g. the coordinated use of defined scales such as the Suffir-Simpson Hurricane Scale for cyclones or the European Macroseismic Scale for earthquakes.

Another issue for future developments is the ability to make use of push notifications. Currently, data is provided mainly without any push mechanisms so that applications have to pull the relevant information creating lots of additional problems such as the emerging data volume that is generated when accessing the data regularly, resulting server performance matters as well as mobile device battery consumption matters generated by each pull access.

Moreover, in TRIDEC Ushahidi is used to validate and demonstrate the feasibility of integrating and using eyewitness reports in an early warning system to make them available to the operators for an improved situation picture in case of tsunamigenic earthquake and tsunami events. Thus Ushahidi provides rapid in-situ crowd-sourced measurement by people actually experiencing the crisis event, e.g. using mobile devices, albeit with variable quality and a high noise to signal ratio which finally is filtered either by automated processing or by manual volunteer efforts. In this way proven and reliable sensors are complemented by human sensors. Whereas sensor data follows standards or undergoes standardisation for format and protocols, the data format of witness reports and the protocols used for sharing them are not subject to standardization activities. This not only affects Ushahidi being the software evangelist in this area, but it also affects imitators and extensions for already established platforms.

As a consequence the use of standards in this context has to be coordinated and should lead to systems whose architecture is flexible and adoptable and complying with common design criteria such as encapsulation of proprietary resources, loose coupling of components, location transparency of services, and separation of concerns. Following these principles it is aimed to realize robust systems which allow interoperability between arbitrary types of systems, seamless integration of various heterogeneous information resources and components, and reuse of components by accessing them via standardized services. De facto standards developed by volunteer communities and developed by consortia with members from industry, academia and authorities have to be considered as well as de jure standards created by official bodies. Standard specifications applied in the context of early warning and crisis management comprise data formats, protocols, and services having in mind existing and possible workflows and knowledge representation.

Discussion including Andy Seaborne [paper] and Herbert Schentz [paper] [slides]

Spatial data is perceived as difficult to work with due to the complexity of the domain. We have taken the approach of providing just simple capabilities to make some handling of spatial data accessible to web developers with little time to become experts in geospatial systems

The long term ecological monitoring and research network (LTER) in Europe[1] provides a vast amount of data with regard to drivers and impacts of environmental change (Mirtl & Krauze 2007, Mirtl et al. 2013). In a test within the EnvEurope (LIFE08 ENV/IT/000399[4]) and ExpeER projects those data have been exposed via a linkedData service and SPARQL endpoint (D2RQ) and an OGC SOS to find out best practices for data services. A SNORQL editor, a SPARQL to EXCEL tool and, most important of all, an R SPARQL plugin have been tested as clients. This access worked quite well and SPARQLing the D2RQ service and the thesaurus SPARQL endpoint proofed feasible.

When using this architecture for accessing distributed services, however, query broker and/or caches are recommendable for performance reasons.

Top requirement, however would be the integration of spatial data and semantics. Although geoSPARQL is already a W3C and OGC recommendation and a geoSPARQL extension to the D2RQ service exists, we could not test that so far because of the lack of geoSPARQL clients.

SPARQLing data sources is nice for the IT-trained domain specialists, our “normal” domain specialist, however would need applications with simple and easy to use GUIS, based on the SPARQL interface.

Lightning Talks: Sensitive Subjects Chair: Titi Roman SINTEF [paper]; Scribe: Andy Seaborne;
(5 mins + 5 mins Q&A each, except 1st one (10 mins))

This position paper introduces DaPaaS – A Data- and Platform-as-a-Service Approach to Efficient Open Data Publication and Consumption. Funded by the European Commission between 2013-2015, DaPaaS aims to simplify data publication for organizations with limited expertise in data publication, and to reduce the cost and complexity related to the data publication and hosting process. At the same time, DaPaaS targets the creation of an ecosystem for development and deployment of third party data-driven applications.

Implementation of the Fukushima radiation LOD framework, Isao Kojima, AIST(National Insititute of Advanced Industrial Science and Technology) [paper] [slides]

In this position paper, we describe about the prototype implementation of LOD based database service for publishing radiation monitoring data of Fukushima area. This database is constructed by converting large number of CSVs to RDF and supports GeoSPARQL.

Open Data-Logging for Bikes, Rich Boakes, University of Portsmouth [paper] [slides]

Cycling in UK cities is often perceived as unsafe, yet there is no useful empirical data recording cyclists experiences. We are working on a project to develop specialised sensor instrumentation for bicycles that will record traffic data automatically, and publish aggregate information, for instance to support analysis by local governments or cycling clubs.

A geospatial API for linking and distributing open data - researching new ways of linking similar objects from different data sets, Bert Spaan, Waag Society [paper] [slides]

The CitySDK Mobility API is a layer-based data distribution and service kit for geospatial Linked Open Data. The API distributes data from different organizations, silos and data sets in a unified way, and lets data owners annotate and add meaning to their data, as well as create links between different data sets. Linking data sets is important when multiple data sets share data about the same object. The API currently creates hard links by connecting two URIs, but this approach is not very flexible: when data can be added, removed or changed from any data sets, and URIs can change, too. This makes keeping links up to date cumbersome and prone to errors. We want to investigate new ways of linking similar objects from different data sets, by using machine learning and by using URIs that do not directly link to objects, but to concepts.

A word about the SmartOpenData project Jesus Mª Estrada Villegas, TRAGSA
Lunch Kindly provided by our hosts, Google
Lightning Talks: More Questions Chair: Linda Van Den Brink; Scribe: Phil Archer; (5 mins + 5 mins Q&A each)
Help! Where next for sustainable statistics and geographies for Wales?, Bill Oates, Welsh Government [paper] [slides]

In this submission, I'm turning the usual conference protocol on its head. I'm not presenting a solution / answer: I'm asking a question.

My team are responsible for the online publication of statistics and geographical datasets for the government of Wales. I need your help and advice in developing a roadmap for data publishing that:

  1. Is evolution rather than revolution for the in-house publishing team’s skill-base;
  2. Can be integrated into mature workflows of the data producers;
  3. Leverages existing investments during their natural lifecycle;
  4. Capitalises on the increasing opportunities to move up the data publishing value chain;
  5. Aims for a combined platform for both statistical and geographical datasets

This may be asking the impossible, but I believe that there are thousands of organisations like mine at a similar level of technical and data publishing maturity. Like those othres, I need your help to evaluate the options in front of me and develop a medium to longer-term road map for my team. I've done some of the groundwork in my paper to produce a list of options and also done a bit of unpacking of the criteria. I'll present these. But where it goes from there is up to you …

How to move a relational Roads Dataset to a Roads Linked Data Dataset?, Alexander Ramage, Transport Scotland [paper] [slides]

Transport Scotland and Agency of the Scottish Government is responsible for the creation and maintenance of the major inter urban roads in Scotland.

The paper identifies the current position and thinking on Linked data in Transport Scotland and identifies those areas that have reasonable understanding and those areas where challenges remain.

The biggest challenge remaining continues to be the Extract Transform and Load moving from a geospatial relational data set to a geospatial Linked Data Set.

Collaborative Geospatial Data, Stuart Harrison, The Open Data Institute [paper] [slides]

Geospatial data is incredibly valuable because it is the underlying data for so much information. It's the glue that can bind together multiple different datasets, and make the dream of a truly linked web of data come true. But the ownership of geospatial data is both distributed and often restricted. OpenStreetMap is an open and collaborative map that may provide part of the answer, but is not part of the linked data web and can be challenging to use in part because of its viral licensing terms. We will outline two collaborative geospatial data projects we are looking at — Open Green Spaces and Open Addresses — and the architectural and governance options that we are examining.

Many layers of thought: connecting INSPIRE and Linked Data, Peter Parslow, Ordnance Survey [slides]

In 1970, Edgar Codd of IBM defined the idea of a relational database, and it quickly became the predominant way to think about data – even though the language of SQL quickly supplanted the original terms. So, ‘tuple’ became ‘row’, ‘attribute’ became ‘column’, and ‘relation’ became ‘table’ – with ‘relation’ itself taking on a different meaning! In practice that ‘relation’ between one tuple/row and another is often mistakenly considered to be the ‘relational’ in relational, because it dominates and constrains the design, being constrained to a hierarchy of parent/child relationships – with common work-arounds.

This approach was quickly adopted in Geographic Information Systems. Geographers had long identified ‘features’ in the world that had interesting properties. Adopting Aristotelian logic (possibly unconsciously!) these features are categorised together – we tend to consider collections of things that have similar properties as instances of the same type of thing. In the 1980s and 1990s, the Open Geospatial Consortium (OGC®) and then the International Organisation for Standardisation (ISO) created a standard ‘meta-model’ for geography. So ‘feature instances’ are stored in rows, with their properties in the columns. ‘Feature types’ are represented in tables: a collection of information about feature instances in the same category.

Meanwhile, in the 1990s, the World Wide Web fell in love with eXtensible Markup Language (XML), with its tree-like data structure. By 2000, the OGC had agreed one of several possible approaches to mapping the relational world they had defined to the tree-like structures being used for data exchange, so Geography Markup Language (GML) was born.

But all that thinking operates at the most abstract level, and at a detailed technical level. The missing part – semantics - was left to each application – what does the data actually mean?

In Europe, INSPIRE represents many years of collective thought about how to classify the geography in a common way. Although imperfect, it is a massive body of agreed semantics.

But now in the 21st Century, along comes ‘linked data’, originally as part of addressing this issue of semantics – a web full of human readable information, but what does any of it mean? Linked data adopts the network or graph model for data management – tuples relate entities to their properties, but they are not forced into tables and constrained to a hierarchy of relationships.

This presentation will give ideas about mapping the layers of thought:

  • Meta-model: how the OGC/ISO General Feature Model can be mapped to the graph model
  • Semantics: how the linked data world could adopt the semantic work of INSPIRE
  • Data Implementation: the easy bit: GML and RDF/XML
  • Service implementation: some ideas on redirection
Representing and Reasoning over Spatial Relations in OWL: A Rule-Based Approach, Sotiris Batsakis, University of Huddersfield [paper] [slides]

Representing spatial information for Semantic Web applications often involves missing or imprecise information. For example the exact coordinates of the boundaries of two regions may be unknown, but it may be known that these regions overlap. This fact can be expressed using qualitative terms such as “Overlapping”. Embedding such information into ontologies and Linked Data is an important practical issue. This paper presents an approach for representing qualitative spatial information and reasoning over such spatial relations. This approach is fully compliant with existing Semantic Web standards and tools. Directions of future work are presented as well.

Panel: Show Me The Way Chair: Stuart Williams, Epimorphics [paper]; Scribe: Hadley Beeman

This topic is offered as potential theme for panel discussion.

With the exception of WGS84 point positions, there is a lack of "dominant common practice" for the storage and expression of geospatial information as linked data. This creates problems for linked data publishers, consumers, application, tool and intrastructure developers/vendors alike.

Alex Coley, DEFRA [paper]

The huge potential of Linked Data is that it will help break down many of the current information silos and add rich and diverse data directly onto the web, gaining and benefiting from the infrastructure of the web itself.

Kerry Taylor, CSIRO

Open Geospatial Consortium and ISO/TC 211 have been developing standards for geospatial data during the last two decades. Two innovations each from OGC and ISO/TC 211 offer useful contributions to the development of a web of linked geospatial data. From OGC these are

  1. the adoption of a uniform URI scheme to identify OGC governed resources and the establishment of a Naming Authority to govern these [1][2];
  2. publication of GeoSPARQL, which extends SPARQL with geospatial functionality [3][4].

From ISO/TC 211 these are

  1. a new (draft) standard providing rules for the development of domain ontologies (application schemas) in OWL2, together with an implementation of the ISO ‘Harmonized Model’ in OWL2 for use in ISO-conformant ontologies [5][6];
  2. establishment of an Ontology Management Group within TC 211 to oversee the publication of ISO-maintained ontologies [7].
  1. Simon Cox (2010), OGC Identifiers - the case for http URIs, OGC Whitepaper 10-124r1
  2. OGC Naming Authority
  3. Matt Perry & John Herring (Eds) (2012) GeoSPARQL – A Geographic Query Language for RDF Data
  5. ISO/DIS 19150-2 (2013) Geographic information -- Ontology -- Part 2: Rules for developing ontologies in the Web Ontology Language (OWL)
  6. Provisional ontologies matching the ISO/TC 211 Harmonized Model
  7. ISO/TC 211 Ontology Management Group
Keith Jeffery, [paper]

An approach to providing linked open data in a geoscience environment is presented. The requirement is for homogeneous access to services, software, resources, users and datasets over heterogeneous provider nodes. The major part of the system is based on advanced heterogeneous distributed database technology using the catalog approach. However, a LOD/semantic web ‘image’ of the system is provided to allow access, browsing and integration in that environment.

Steve Peters, [paper]

The purpose of this paper is to stimulate a discussion on:

  • Potential improvements to the GeoSPARQL standards, to support more sophisticated management and re-use of geographic LinkedData.
  • Future relationships between and the respective roles of triple-stores and spatially-enabled databases as technologies to power more open, integrated geo-spatial data and intelligence.

Our main observations and conclusions are that:

  • LinkedData and SPARQL technologies are increasingly being adopted by Government organisations as channels for routinely disseminating significant data assets – examples include: DEFRA’s sets on Bathing Water quality; Land Registry’s Price Paid series; geographic data from Ordnance Survey and ONS, and ; DCLG’s “Open Data Communities” service.
  • Geography and “place” are key to unlocking and binding together datasets within and across these services.
  • GeoSPARQL therefore has an important enabling role: it could be the foundation for triple-stores as products akin to the spatially-enabled Relational Databases powering the GeoWeb.


  • Current GeoSPARQL standards appear to be limited to straightforward spatial querying – e.g. to identify “points in polygons”, or touching/overlapping geometries. There are emerging requirements to support more sophisticated use cases
  • Take-up of GeoSPARQL in commercial and open source triple-stores appears to be very limited. Further work is needed to understand why, and identify how to improve take-up across the vendor community.
  • In our experience, it can be difficult for users to locate and consume geospatial LinkedData in their application. We need better tooling to support geo-LinkedData reuse in a range of scenarios – including desk-top GIS, and web-based mapping services such as OpenStreetMap or CartoDB.
  • Alternative geo-spatially enabled databases are well established as both commercial - e.g. Oracle - and open-source - e.g. PostGIS - products, and have a strong track-record in delivering geographic data using open standards.
  • A key question is: should triple-stores, enabled by GeoSPARQL, evolve to provide features and functionality akin to a geo-spatial database? Our view is yes – not least because it helps to standardise and streamline management of LinkedData assets, plus creates opportunities for innovative new analysis using datasets in the Semantic-Web. However, the reality is that geographic data publishers will continue to operate hybrid triple-stores/ geospatial database solutions for some time. Further work is therefore needed to strengthen integration between these two technologies.
John Goodwin, Ordnance Survey [GeoSPARQL]
Raphaël Troncy, Eurecom [NeoGeo spec]
Bar Camp Pitches Moderator: Bart de Lathouwer
Pitch your discussion idea in 60 seconds or less! Ideas already notified:
  • Time-dynamic contextualized POIs, Steven Verstockt, iMINDS
  • How can we generically subset an already selected big blob of data? Chris Little, Met Office
  • Integrating linked spatial data with the Web: proposal to promote native support in HTML vocabulary with REST and media types, Peter Rushforth, Canada Center for Mapping and Earth Observation
  • Using W3C Open Annotation to provide links between metadata and geotemporal subsets of EO datasets, Raquel Alegre, University of Reading
  • Query and Visualisation of SDI and Linked Open Geo Data in common Web client interfaces, Martin Tuchyňa, SAŽP (Slovak Environmental Agency)
  • Promoting a GeoJSON-LD crossover 'standard', a format for bridging spatial and semantic front end application components and uses, Rein van 't Veer, Digitaal Erfgoed Nederland/Erfgoed en Locatie
  • Metadata for predicates, Peter Parslow, Ordnance Survey
  • Leveraging of crowdsourced environment information, Gianfranco Gliozzo, Extreme Citizen Science (ExCiteS) research group
  • Linked data and OGC/other standards - existing and new - can they facilitate the integration of non-authoritative data in decision-making processes? Bente Lilja Bye, BLB
  • Determining which RDF ontologies provide best practices in areas of overlap, Jeremy Tandy, Met Office
  • Fuzzy URIs, Bart Spaan, Waag Society
  • How shall we evolve the Raspberry Pi prototype into a project that is useful to as many of us as possible? Rich Boakes, Portsmouth University
Bar Camp
Find your coffee on the third floor
Bar Camp reports & wrap up Chair: Phil Archer Scribe: Someone from each group;

Hash Tag

The hashtag for the event is #lgd14. Tweets with this hash tag may be quoted in the report of this event.

IRC Channel

Make comments, ask questions, follow along, help to record the procedings by joining the chat room.

Short version: use the Web interface and join channel #lgd

Longer version, port numbers etc. See W3C IRC page.

Campus logistics

The workshop is on the ground floor, the café is in the basement.

There is strong wi-fi throughout the building.


Smart Open Data logo
W3C is acting on behalf of the Smart Open Data project, in which it is a partner, co-funded by the European Commission under the Seventh Framework Programme

The event is co-organized by the UK Government, Ordnance Survey, the OGC and Google.


W3C gratefully acknowledges Google for hosting this workshop.


Important Dates


Expression of interest — please send a short e-mail to Phil Archer ASAP.

19 January 2014:
Deadline for Position Papers
(EasyChair submission)

5 February 2014:
Acceptance notification and registration instructions sent. Program and position papers posted on the workshop website.

5th March 2014, 09:00
Workshop begins

19:00 - 22:00 Recommended evening event: London Open Data Meetup at the Open Knowledge Foundation

6th March 17:00
Workshop ends