As you probably have noticed, for a couple of years now there has been a trend towards building smart cities; that is, cities embracing new technologies like cloud, mobility and the Internet of Things. All these technologies directly impact the size and speed at which data becomes available. It is this data, or Big Data, together with already available data, which is currently being explored as the means to optimise city operations, to create a better citizen experience and facilitate collaboration.
At the same time this data should also accommodate the growing migration of people into cities, create a better living and working environment and reduce the impact on the natural environment. Many of the projects in these smart cities initiatives around the world, however, are currently experiencing some challenges. The challenges are not to do with technology, but with information interoperability and making sense out of the data available. Issues that are becoming increasingly relevant to all types of modern businesses.
Regardless of whether the data is coming from a sensor or out of a database, it needs meaning (what) and context (who, where and when) to make sense out of it, so it can become useful information. As this mix of data needs to be interoperable, standards are becoming crucial to the integration of data sets.
Unfortunately, experiences from the past have taught us that data stored in silos, with the meaning of data embedded in applications, makes it virtually impossible to make this data transparently available, beyond the scope of the application, without very complex migration projects.
Data in a smart cities context ranges from pollution data coming from sensors, through behaviour patterns from social networking and mobile devices, to authoritative data about the cities infrastructure and built environment.
Smart cities projects therefore tend to be complex, involving many different organisations such as local government, utilities and insurers, as well as complex data sets (location, electoral roll). Smart cities also involve complex spatial data and the integration of this data with other sets. Those involved in spatial data however have already worked out ways to convert this Big Data into something useful to the business. This is because spatial data tends to get very big, very quickly. It not only has to intersect with the data below (underground system), above (roof gardens) but adjacent too (hotel, office, home).
So standards (often semantic) are becoming increasingly important. Data sets need to be integrated and organisations need technology to be able to deal with this. Spatial data can be very complicated, but the lessons of smart cities projects can be passed on to businesses in general.
Currently a lot of standardisation efforts are evolving around semantics and business rules that data has to comply to, in order to be able to deal with transparent information interoperability.
The most important ones we see are at the World Wide Web Consortium (W3C) on Ontologies (OWL), Vocabularies (RDFS), Rules (SWRL), Rule Exchange (RIF) and the RDF query language (Geo)SPARQL (like SQL for relational databases) and the Open Spatial Consortium in conjunction with ISO specifically for the Spatial Domain who are currently aligning with the W3C in the Spatial Data on the Web Workgroup as part of the GeoSemantics Domain Workgroup.
These standards are still evolving but we should be careful not to diverge instead of converge because too many standards won’t work either.
Nevertheless, because of the requirements for information interoperability, smart city reference models like the Smart City Data Concept Model as defined in PAS182 by BSI in the UK (top down) or the System of Key Registers in the Netherlands (bottom up), are evolving to allow for flexible inclusion of specific domain models of e.g. utility infrastructure or land administration into higher level models. Choosing appropriate semantic and rule standards independently from the application logic are indispensable for information interoperability.
The latest development around semantic interoperability is linked data on the web as defined by W3C, which is based on RDF but is finding an alternative in (Geo)JSON-LD, which comes out the REST/ JSON (versus SOAP/XML) paradigm. Regardless of which paradigm is chosen, semantics plays a key role and allows for information interoperability and web service interoperability without the need for data migration, as the integration is done on a metadata level. Currently some organisations are investigating whether the linked data approach might be an alternative for their existing data integration efforts towards enterprise information interoperability.
Privacy and security
Ultimately new technology will find its way to better secure data, especially when we have a better insight into our data. If we better understand what the meaning is and in which context it is relevant independently of any application logic, we also have better ways to isolate it, which is particularly true if dealing with volatile and/or private data. The linked data approach might actually help evolve this, as only through the metadata is data accessible, while data in itself can be meaningless without metadata. Although many algorithms do exist in data analytics tools which allow for detecting patterns in data, especially when data from different sources are brought together even when anonymous, which could lead to very specific information about someone at a specific place and time.
Data quality and governance
Although an important issue for modern businesses, data quality and governance is still a very unpopular subject. Perhaps it has to do with the data silos we have been building for the last 30 years with no semantics and no rules to keep the data clean and up to date, or maybe organisations are just not aware of these issues and how important they are. It could therefore take some significant effort to change how things are done.
Even in areas that have had very good experiences applying semantics and rules, such as in the spatial domain sector, we are still seeing a lot of reluctance in adopting adequate open standards based tooling.
The benefits of adopting semantics and rules and having a data quality and governance strategy are actually quite obvious; bad data leads to bad or no decisions being made, especially when a mix of data sources is required. This applies to smart cities wanting to improve quality of life, as well as businesses, which are also increasingly dependent on both external and internal data to make important decisions. It is essentially an issue about information interoperability.
By Han Wammes, Business Development Manager at 1Spatial