We introduce the concept of City Service Ecosystem (CSE) as digital environment for the governance of urban services. We trace the research challenges and opportunities of adding semantics to improve the management of such ecosystems, especially in relation to description and retrieval of urban-related Web services. We explain the peculiarities and distinct characteristics of CSEs resulting from their relation to the city space and we introduce our current work to enhance with semantics an existing CSE in the city of Milano, in relation to the forthcoming World Exposition (EXPO 2015).
Cities are complex environments in which multiple heterogeneous stakeholders produce, consume and exchange digital information. Public utilities monitor their goods consumption (water, electricity, waste, etc.) and gather information from complementary sources to improve their business and service performance; public bodies provide information and services to their citizens with the aim to satisfy people needs and with the mandate to govern and coordinate the actions on the territory; citizens request and provide information about what happens in their cities during their everyday activities as part of their lives.
It is thus clear that in every city a diverse and intertwined ecosystem of
services exists and that multiple and independent information systems are put
in place to manage the digital representation of those services. We name City
Service Ecosystem (CSE) the digital environment in which those services live,
according to the principles and guidelines of the so-called API economy [
The peculiarity of a CSE with regards to other digital environments (e.g. any other B2B system) lies in the strong connection with a spatio-temporal context : those services are related and insist on a physical area and are widely dependent on the timing and possible co-occurrences of events happening in the city.
Taking the example of city mobility, multiple actors o er information services: local transportation utilities (timetables of buses and subways, availability of taxis), long-range transport companies (train and ights arrivals and departures), environmental-friendly or sustainable mobility means (availability of bike sharing or car sharing) and of course private transportation (GPS traces of vehicles, tra c sensors in the streets). The mobility providers are in a business competition, but, if they cooperate by exchanging their information (passenger ows, availability of services, congestion and delays), the whole city ecosystem bene ts and the quality of life of citizens improves.
Our investigation questions are therefore: how should such a CSE be shaped and organized? How can semantics support the life-cycle of city services? This short paper is aimed at exploring (a part of) such a research space and at providing preliminary answers to those questions. 3
Managing city services within a CSE means managing their descriptions in the rst place. In this section, we address the issue of using semantics to make explicit and machine-readable their speci cations. Beyond the general considerations applicable to the semantic description of any Web service, we focus on the speci city of urban services. 3.1
There is a large literature about ontological models and semantic enrichment of SOAP-based Web services [3{6]. However, those proposals have shown limited adoption outside the research community; one of the reasons is probably the high cost of producing such descriptions, making it an entry barrier that discouraged a real uptake.
Another reason is the rising popularity of Web APIs and REST services [
Since cities are complex environments, their services span a large set of di
erent domains, covering the heterogeneous aspects of citizens' lives. Nonetheless,
they all share the city context from both a spatial and temporal point of view
[
Adopting common models to describe geographic and temporal features, as well as building and re-using identi ers for relevant points of interest (POI) in a city becomes therefore of utmost importance to add the relevant semantics to urban services. The question arises whether spatio-temporal semantics should be used to annotate the description of a service interface rather than to \shape" the data exchanged through the services, so to enable the interoperability of services' requests and responses. We think that both those conditions apply: in the former case, semantics supports the selection of the relevant service on the basis of the requester's needs (e.g., choosing the best transportation service for a speci c geographical area); in the latter case, semantics allows an easier mash-up of di erent services in the same context (e.g., combining two di erent transportation services on the basis of the proximity between the stops/stations of the di erent providers so to enable a multi-modal journey planning). 3.2
Web service descriptions are intended in the rst place to be employed to discover
{ either at design time or at run-time { the best service to ful l a speci c need.
Research focused also on semantically-enhanced Web service discovery
(including our own results [
We believe that a semi-automatic or only partially automated scenario is
more likely to happen: a developer needs a service to complete a business
process and would like to nd it, exactly as he would do by searching the Web for
information. As a consequence, in this paper we talk about Web service retrieval
rather than discovery. A number of Web service catalogues and directories have
1 A very interesting discussion about the interplay between spatial and temporal
description is taking place within the W3C Locations and Addresses Community
Group, cf. http://www.w3.org/community/locadd/.
emerged2, both as pure editorial e orts or collaborative wiki-like initiatives;
developers can look for services via traditional keyword-based search or category
facets/ lters. Along those lines, some proposals emerged within the Semantic
Web community to build scalable \search engines" for Web services [
In the context of cities, the same considerations are valid and it is clear that a city service description should include spatial information to allow for a geo-based retrieval (e.g. which public transportation services cover an area of x kilometre radius from this position?). Within a CSE, in which multiple actors cooperate and compete, other service details are very important, like usage terms and conditions: a semantic description of such non-functional properties of urban services should be mandatory for a semantic CSE.
It is worth noting that such non-functional description are not only limited to generic business rules (e.g., whether a service is available to everybody vs. limited to registered users, or provided for free vs. requiring a fee), but also to city-speci c peculiarities and conditions related to the actual usage of such services (e.g. is a ticket from a transportation service valid to travel with another provider? will a shuttle bus connection service wait for the arrival of a speci c train service in case of delay?): with the increasing blending of the physical and cyber systems, digital representations of such features becomes essential for an e ective governance of a CSE. 4
Milan will host the World Exposition in 2015 whose main theme is \Feeding the Planet, Energy for Life"3. This event is a major opportunity to introduce disruptive innovations in all aspects of the urban life: infrastructures, transportation, cultural and social life, accommodation, services and facilities.
E 15 [
While in its current realization, E 15 does not make use of semantic technologies, all the issues related to an e ective description of the o ered services emerged during its design and development. In order to lower the entry barrier 2 Like the very popular ProgrammableWeb, http://www.programmableweb.com/. 3 Cf. http://en.expo2015.org/. 4 Cf. http://www.e015.expo2015.org/. for participants, E 15 requires providers to compile a form to describe their services (besides producing the technical details, for example WSDL les): the completed forms can be the basis for a semantic description of E 15 services. We are currently exploring and comparing the di erent existing ontologies to describe Web services, in order to come up with a simple but powerful model, able to cover both SOAP-based services and APIs and REST services; we are adopting a bottom-up approach, by starting from the concrete E 15 services so to take into consideration all the concrete requirements coming from a real CSE. While E 15 currently encompasses services mainly in the domains of transportation, accommodation and tourism, we believe it is an interesting test-bed to experiment with semantic technologies.
E 15 also fosters interoperability at data level by promoting sharing and reuse of glossaries between di erent services; for example, the services providing real-time information about ight arrivals and departures of the three airports around Milano, even if operated by di erent stakeholders, adopted the same service interface and data structures5. Semantic technologies can for sure play an important role in de ning common and shared vocabularies and ontologies and in paving the ground for an improved interoperability of city services. 5
In this paper, we introduced the concept of a City Service Ecosystem and we explained the rational and the possible bene ts of adopting \a little semantics"6 to improve city service descriptions and enable an easier service retrieval and an improved interoperability at city level. We are currently applying the considerations expressed in this paper to enhance the E 15 service ecosystem of Milano; our plan is to design and develop a semantic registry for city services (something like LOV7 for services) and to enable a semantically-enhanced retrieval of those services.
This work was supported by SPAC3 (project id 40696059) co-funded by Regione Lombardia (POR-FESR 2007-2013), and by 3cixty (id 14523) and Connecting Digital Cities (id 14465) projects co-funded by EIT. 5 The di erent providers were invited to agree on a common conceptualization and they thus adopted the same model, perceiving the value of sharing such \semantics". 6 Cf. http://www.cs.rpi.edu/~hendler/LittleSemanticsWeb.html. 7 CF. http://lov.okfn.org/.