News Explore the world of Geo-information

Frans Knibbe

Spatial data on the Web: goals of the OGC/W3C working group

8 February 2016

In 2014 the OGC and the W3C joined forces to make it easier to publish and use spatial data on the worldwide web. The Spatial Data on the Web Working Group was formed to improve and clarify the necessary standards, something that is very much needed.

In my previous blog I wrote about the nature of spatial data, the importance of having spatial data on the web and about the current transition phase that Geo-information finds itself in – data no longer have to reside in closed-off silos but can be shared and combined freely. The OGC and W3C, the two major standards organizations for Geo-information and the worldwide web respectively, understand that guidance is needed for this transition to go smooth. So a joint working group, the Spatial Data on the Web Working Group (SDWWG) was formed. Never before had the two consortia worked together this way – it took some time to find a mode of collaboration that both felt comfortable with. But they did it, which is a minor success by itself.

The formal document that sets out collaboration in the working group is its charter. It contains a description of the mission and deliverables of the group. In this blog I want to discuss those goals (in the next blog I will write about how the goals can be accomplished).

The mission

The working group aims to improve and clarify the standards needed for dealing with spatial data on the web. This is a noble and useful mission, because current standards do not form a neatly organised collection. In some areas there are too many standards, for example about how a vector geometry should be encoded. In other areas there is an absence of standards, on how to define and use coordinate reference systems for example. Also, standards do not always connect well to each other.

The current messiness of standards can be partly explained by the omnipresence of spatial data. The large majority of data we use every day has spatial aspects. This means that whoever manages or publishes those data will have to deal with their spatialness. Many people and organizations have been involved in standardization of spatial data, often from a certain domain perspective and sometimes before the web was recognized as a viable platform for data exchange. The OGC, for example, has created an extensive collection of standards for geographic data, forming an interdependent and complex closed semantic system. Other domains too have found it necessary to develop standards for common types of spatial data, like toponyms, addresses or point locations. The web, however, is a place where all domains meet, so domain standards won’t do. It needs general standards that are as simple as possible.

What to expect from the Working Group?

The charter neatly lists everything that the Working Group intends to deliver:

  1. A document with use cases and requirements for spatial data on the web;
  2. An overview of best practices for publishing or consuming spatial data on the web;
  3. An improved version of the Time Ontology in OWL, a standard for expressing temporal data on the web;
  4. An improved version of the Semantic Sensor Network Ontology, a standard for expressing sensor data on the web;
  5. A recommended way (presumably: a vocabulary) for expressing coverage data on the web

These results will be published both by the OGC and the W3C, with the last three deliverables having a more official status than the first two. Each of the deliverables is interesting in its own way:

Use cases as a basis

Requirements for working with spatial data on the web were largely derived from collected use cases. Because the requirements establish what should be done, the Working Group made creating the Use Cases and Requirements document its first priority. The document contains a fair number of different ways people want to use spatial data on the web. In those narratives common elements were identified, leading to a number of short statements on how things should be – the requirements for the other deliverables. A few important requirements that are identified are:

  • Clear-cut standards for expressing (encoding) spatial data, spatial metadata and spatial relations;
  • Allow spatial data to be linked to other (spatial) data;
  • Allow searching and finding spatial data on the web.

Best Practices

Currently the Working Group is working on a second document that will contain recommendations on how to publish, find and use spatial data on the web. The document should be understandable for a broad audience, certainly not only for experts in the fields of spatial data, or the web. As much as possible, the recommendations will be based on evidence found in the field and on general practices for handling web data (e.g. using JSON and RESTful APIs).

The Working Group will therefore inventory current standards, tools and practices. In cases where functional overlap is found choices will have to be made. Functional gaps can be discovered too: problem areas for which there is no applicable solution. In those cases the Working Group will have to try to offer a solution, which could take the shape of extending existing web semantics in cooperation with other communities that have shared goals.

Temporal data

At first sight it may seem strange that a working group about spatial data will occupy itself with the subject of time. After all, on a non-cosmic level, space and time can be dealt with separately: data can have spatial components, temporal components, or both. Still there is some sense in the decision to take time on board:

  1. Standards for publishing or using spatial data on the web at the moment do often not suffice for serious, professional data applications by the likes of governmental organizations or research institutes. Those sort of data often have a temporal aspect, for instance because the things that the data are about change over time.
  2. Sharing spatial data is important in sciences like history, archeology and geology. One of the intended improvements of the Time Ontology is improving support for time that can not be expressed in the gregorian calendar or ISO 8601. Especially in sciences that deal with the past such functionality is needed.
  3. The OGC has already modelled the concept of time in several standards for geospatial information, for example for time series, which are considered a type of coverage and thus a type of geospatial data.
  4. There is much logical overlap between space and time. For instance, this can be seen by comparing Allen’s interval algebra to the DE-9IM. Similarities between time and one dimensional space, like it is used in linear referencing, are also striking. So there seem to be possibilities for models of space and time that are harmonized in some respects.
  5. Our planet Earth is far from static, it keeps changing in many ways. For example, tectonic plates keep moving and changing their relative positions. This creates a time dependency in geographical coordinate reference systems.

I do not know if any of these reasons were decisive in putting time within the scope of the Working Group, but I do know it is very useful to be able to use such a common and important concept on the web of data.

Sensor data

Sensors, devices that measure something in their environment, have two spatial aspects. Firstly, they have a location (which can be fixed or variable). Secondly, sensors often measure a spatial phenomenon, like air pollution or the location of people or objects. So it is understandable that the OGC concerns itself with the development of standards for sensor data. And, since 2011, there is a standard for the data web, the Semantic Sensor Network (SSN). It features a usable ontology, but one that could be improved to finish the work of the SSN Incubator Group. Improving possibilities and ease of use are among the targets.

This deliverable should contribute to general availability, interoperability and workability of sensor data on the web. And that will offer many opportunities for the Internet (or Web) of Things and for smart environments and smart infrastructures.

Coverage data

A coverage is a collection of data that describe a phenomenon that is variable in time and/or space. It can cover one to four dimensions. Examples are satellite images, tomograms, time series and point clouds. As such, it is an important class of spatial data, especially in research. Think of climate or environmental research for example.

The work on this theme will likely result in an ontology for coverage data on the web. It should be based on existing and developing OGC specifications, but could also build upon the Data Cube Vocabulary, a web ontology for multidimensional data.

Ambitious. Perhaps too ambitious?

Taken together, what the Working Group intends to achieve can be called very ambitious. Should the group succeed in achieving its goals, the web will certainly experience a tremendous boost in its data sharing abilities. But more isolated connected systems like organization intranets can also benefit from solutions for making data and systems more linkable.

Are all the set goals really achievable? The group consists of almost seventy experts on spatial data and/or web data, but for most of them involvement in the group is not a full time employment but a voluntary contribution, next to many other pressing everyday duties. Also the difficulty of some of the problems that the group faces should not be underestimated. Time will tell, but I have hope for very useful results around the end of 2016.

This hope is fueled by the pragmatism of the group. The wheel will not be reinvented, careful use is made of existing specifications and analyses. Also the working group strives to cooperate and share the load with other communities. Besides that, the worldwide web is a platform made up from modules that can be smartly (re)combined for self-improvement and growth. So even if the Spatial Data on the Web Working Group does not meet all of its goals, its results can certainly be put to good use to continuing building a better web of data for the world.

In the next blog of this series I will discuss some of the issues requiring better standardization, in order to accomplish a web in which spatial data can really come into bloom.