The small city of Cluj in Romania, of some half-million inhabitants, is responsible for a 2.5 million triple store, as part of a Recognos-led project to develop a “Linked City” community portal. The project was submitted for this year’s ICT Call – SME initiative on Digital Content and Languages, FP7-ICT-2011-SME-DCL. While it didn’t receive funding from that competition, Recognos semantic web researcher Dia Miron, is hopeful of securing help from alternate sources in the coming year to expand the project, including potentially bringing the concept of linked cities to other communities in Romania or elsewhere in Europe.
The idea was to publish information from sources such as local businesses about their services and products, as well as data related to the local government and city events, points of interest and projects, using the Linked Data paradigm, says Miron. Data would also be geolocated. “So we take all the information we can get about a city so that people can exploit it in a uniform manner,” she says.
The first step was to gather the data and publish it in a standard format using RDF and OWL; the next phase, which hasn’t taken place yet (it’s funding-dependent), is to build exportation channels for the data. “First we wanted a simple query engine that will exploit the data, and then we wanted to build a faceted search mechanism for those who don’t know the data structure to exploit and navigate through the data,” she says. “We wanted to make it easier for someone not very acquainted with the models. Then we wanted also to provide some kind of SMS querying because people may not always be at their desks. And also the final query service was an augmented reality application to be used to explore the city or to navigate through the city to points of interest or business locations.”
Local Cluj authorities don’t have the budgets to support the continuation of the project on their own, but Miron says the applications will be very generic and can easily be transferred to support other cities, if they’re interested in helping to support the effort. Other collaborators on the project include Ontotext and STI Innsbruck, as well as the local Cluj council.
I don’t doubt this would be useful information for users but is this the delivery model that is going to work for users, assuming it is funded? Here or elsewhere?
How hard do users work with searches? See Keyword and Search Engines Statistics to get an idea by country.
Some users can be trained to perform fairly complex searches but I suspect that is a distinct minority. And the type of searches that need to be performed vary by domain.
For example, earlier today, I was searching for information on “spectral graph theory,” which I suspect has different requirements than searching for 24-hour sushi bars within a given geographic area.
I am not sure how to isolate those different requirements, much less test how close any approach is to satisfying them, but I do think both areas merit serious investigation.