semantics for smarter cities

LD4SC-‐15

Seman&cs for Smarter Ci&es. Hype or Reality? A Study in Dublin, Bologna, Miami, and Rio

Freddy Lecue

1

LD4SC-‐15

Context

LD4SC-‐15

Arrival of the “urban millennium” By 2050 over 6 billion people, two thirds of humanity, will be living in towns and cities

Source: Smart Cities How will we manage our cities in the 21C?, Colin Harrison IBM Corporate Strategy Member, IBM Academy of Technology

Indoor air pollution resulting from the use of solid fuels [by poorer segments of society] is a major killer! Claims the lives of 1.5 million people each year, more than half of them below the age of five (4000 deaths per day)

Water problems affect half of humanity!!!! 1.1 billion people in developing countries have inadequate access to water, and 2.6 billion lack basic sanitation

1.6 billion people — a quarter of humanity — live without electricity! South Asia, Sub-Saharan Africa and East Asia have the greatest number of people living without electricity (as high as 706 million in South Asia) Source: 2008 UN Habitat; Smart Cities How will we manage our

cities in the 21C?, Colin Harrison IBM Corporate Strategy

Introduc&on Mo&va&on – Today’s Ci&es are Confronted with Serious Dilemmas

LD4SC-‐15

ASIA N.A. & EUROPE AFRICA LATIN AMERICA

Rapid expansion Negative growth Rural exodus increasing poverty Decentralization

§  Over the next decade, Asia’s urban areas will grow by more than 100,000 people a day §  Growth rates are more rapid than the investment in infrastructure § Benefits of new infrastructure investments have not been distributed equally

§  46 countries (including Germany, Italy, most former Soviet states) are expected to be smaller in 2050 §  The number of shrinking cities has increased faster in the last 50 years than the number of expanding cities

§  In 2008, more than 12M Africans left their rural homes to live in urban areas §  The projected increase in urban migration will exacerbate the problems of providing infrastructure, sanitation. health services, and food

§  Large cities have incorporated nearby villages and towns – as a result, large urban areas developed sub-centers whose functions duplicated those of the central city §  Many large cities are competing with their outlying suburbs for people, revenue, and employment

Source: Various; IBM MI Analysis;

Introduc&on Mo&va&on – Regions have both Common and Unique Challenges

LD4SC-‐15

Source: Various; IBM MI Analysis;

Introduc&on

The Sustainable Eco-City The Well Planned City The Healthy and Safe City

The Cultural-Convention Hub The City of Digital Innovation

Mo&va&on – Beyond the prac&cal objec&ves, ci&es have ‘aspira&ons’

The City of Commerce

LD4SC-‐15

Miami, USA

Dublin, Ireland

Rio, Brazil

• 5.5 billion hours of travel delay • 2.9 billion gallons of wasted fuel in the USA • $121 billion / year • 0.7% of USA GDP *5 over the past 30 years

Bologna, Italy

How to reduce traffic congesHon ?

Introduc&on Mo&va&on – Socio-‐Economic Context

LD4SC-‐15

Introduc&on Mo&va&on – Limita&on of Exis&ng Systems (Example: Traffic) (1)

Most traffic systems

already support Basic Analy&cs

and Visualiza&on!!!

Basic AnalyHcs from Bus / Taxi data

LD4SC-‐15

•  All exis&ng traffic management systems are based on ONE signal / stream •  No possible interpretaHon of traffic Anomalies •  No IntegraHon of Exogenous Data

No SemanHcs / Context !

No ExplanaHon No Diagnosis

Introduc&on Mo&va&on – Limita&on of Exis&ng Systems (Example: Traffic) (2)

LD4SC-‐15

•  Sensor data assimilaHon –  Data diversity, heterogeneity –  Data accuracy, sparsity –  Data volume

•  Modelling human demand

–  Understand how people use the city infrastructure –  Infer demand pa_erns

•  Factor in Uncertainty –  Opera&ons and planning –  Organise and open data and knowledge, to engage

ci&zens, empower universi&es and enable business

Introduc&on Seman&cs for ci&es: How can Seman&cs help ci&es transform ?

LD4SC-‐15

Introduc&on Seman&cs for ci&es: why now? Open Data!

LD4SC-‐15

Introduc&on Seman&cs for ci&es: why now? Ci&es want to be Smarter Ci&es!

Nowadays: Issues of Urban Quality such as housing, economy, culture, social and environmental condi&ons.

LD4SC-‐15

Introduc&on Seman&cs for ci&es: why Seman&cs? Big Data

LD4SC-‐15

Working harder is not sustainable

Cities require innovative approaches

Introduc&on Seman&cs for ci&es: why Seman&c Web, Linked Open Data? Smart Systems

LD4SC-‐15

Introduc&on Seman&cs for ci&es

Walkonomics

FixMyStreet

Schooloscope

FloodAlerts

BeatTheBurglar

Where can I live?

LD4SC-‐15

Introduc&on Seman&cs for ci&es – Our Experience

LD4SC-‐15

Data format and data access, collec&on, storage, transforma&on Big Data – The World of Data Source: Various; IBM MI Analysis;

LD4SC-‐15

Data format and data access, collec&on, storage, transforma&on Big Data – 4Vs Source: Various; IBM MI Analysis;

LD4SC-‐15

Data format and data access, collec&on, storage, transforma&on Data Format and Heterogeneity -‐ City Data != Open Data (1)

A�Global�Movement�Has�Begun�to�Provide�Transparency�and�Democratization�

of�Data�

18November�30,�2011

Don’t�see�your�site?�Update�via�@usdatagov

(*) “Driving Innova&on with Open Data”, Jeanne Holm, Data.gov, February 9th, 2012 (Presenta&on to Ontology 2012)

Think�Big,�Start�Small,�InnovateData.gov�Quick�Facts May�2009 October�2011

Total�datasets�available 47 >400,000

Hits�to�Data.gov 0 >200�million

Apps�and�mashͲups�by�citizens�and�government 0 372�+�1113

RDF�triples�for�semantic�applications 0 6.7�billion

Dataset�downloads 0 >2.0�million

Nations�establishing�open�data�sites 0 28

States�offering�open�data�sites 0 31

Cities�in�North�America�with�open�data�sites 0 13

Open�data�contacts�in�Federal�agencies 24 396

Agencies�and�subagencies�participating 7 185

Communities 0 7

Community�challenges 0 23

November�30,�2011 11

A lot of relevant open data for city data analy&cs

LD4SC-‐15

Applica&ons in Ci&es: Bologna, Dublin, Miami, Rio… and

Airlines

LD4SC-‐15

Miami, USA

STAR-‐CITY (Seman&c Traffic Analy&cs and Reasoning for CITY)

Input

•  Type of anomaly (bus speed, ambulance delay…)

•  Type of explana&on (city events, unplanned events, road works, …)

Output

•  Impact of events and their characteris&cs on anomalies

20

SemanHc InterpretaHon of

Diagnosis

SemanHc Reasoning: Diagnosis

SemanHc Context

SemanHc Search: Diagnosis ExploraHon

Live IBM demo: h_p://208.43.99.116:9080/simplicity/demo.jsp Live WWW demo: h_p://dublinked.ie/sandbox/star-‐city/ Video: h_p://goo.gl/TuwNyL

LD4SC-‐15

STAR-‐CITY (Seman&c Traffic Analy&cs and Reasoning for CITY) Context •  Dublin: (Diagnosis of) Traffic conges&on

•  Bologna: (Diagnosis of) Bus conges&on •  Miami: (Diagnosis of) Bus bunching •  Rio: (Diagnosis of) Low on-‐&me performance of buses

Source of Anomaly

Source of Diagnosis

LD4SC-‐15

Challenge:

Source of Anomaly

Source of Diagnosis

•  Explaining Traffic Condi&ons with Diagnosis Reasoning

•  Logical correla&on of anomalies and diagnosis in dynamic sepngs

•  Knowledge Representa&on and Reasoning •  Machine Learning / AI Diagnosis •  Database: Large scale data integra&on •  Signal Processing / Stream Reasoning

Core Areas / Problems:


LD4SC-‐15

Architecture of Diagnosis Components


LD4SC-‐15

STAR-‐CITY (Seman&c Traffic Analy&cs and Reasoning for CITY) Data

Vocabulary

LD4SC-‐15

Challenge: Predic&ve reasoning (as opposed to analyHcs) in heterogeneous and dynamic sepngs

•  Knowledge Representa&on and Reasoning •  Machine Learning / Knowledge Discovery •  Database: Large scale data integra&on •  Signal Processing / Stream Reasoning

Core Areas / Problems:

Traffic Condi&on

Road Incident

Weather Condi&on


LD4SC-‐15

Miami, USA


Input



Output


Input



Output

•  Real-‐&me/Historical diagnosis results

•  Evolu&on of diagnosis over &me/space

•  Comparison vs. historical

Objec&ve: Real-‐Time and Historical Traffic Diagnosis (1) Live demo: h_p://208.43.99.116:9080/simplicity/demo.jsp Video: h_p://goo.gl/TuwNyL

LD4SC-‐15

Miami, USA


Input



Output


Input



Output

•  Categoriza&on of diagnosis

Live demo: h_p://208.43.99.116:9080/simplicity/demo.jsp Video: h_p://goo.gl/TuwNyL

Objec&ve: Real-‐Time and Historical Traffic Diagnosis (2)

LD4SC-‐15

Miami, USA h_p://vtce.altervista.org/

Reverse STAR-‐CITY system for city managers

LD4SC-‐15

Reverse STAR-‐CITY system for city managers Objec&ve: City Planning

Input



Output


h_p://www.vtce.altervista.org/

LD4SC-‐15

Reverse STAR-‐CITY system for city managers Data

Vocabulary

LD4SC-‐15

Miami, USA

Mobile STAR-‐CITY app for ci&zen in Dublin and Bologna

LD4SC-‐15

Miami, USA


Use case scenario: MeeHng the Increased Mobility Demand: •  Scenario 1 “Road Traffic Diagnosis” •  Scenario 2 “Road Traffic Predic&on” •  Scenario 3 “Personalized Traffic Restric&ons”

Outcome: •  One mobile app •  Two pilot ci&es (Dublin and Bologna) •  Live and real-‐&me environment •  Real data (user calendar, open data: traffic

conges&on, weather, events, road works, accident …) •  but with simulated car sensor data – Aus&n ;-‐)

Context

LD4SC-‐15

Mobile STAR-‐CITY app for ci&zen in Dublin and Bologna Data

Dublin

Car-‐related

User-‐related

Vocabulary

Bologna

LD4SC-‐15


User Context-‐aware Driving (User data)

Open Context-‐aware Driving (Open data)

Private Context-‐aware Driving (City data)

Objec&ve: Context-‐aware driving experience (1)

LD4SC-‐15

Mobile STAR-‐CITY app for ci&zen in Dublin and Bologna Objec&ve: Context-‐aware driving experience (2)

Personal Events-‐aware Driving Car Sensor

Simula@on

Car Sensor-‐aware Driving

LD4SC-‐15

Real-‐Time Urban Monitoring in Dublin

h_ps://www.youtube.com/watch?v=ImTI0jm3OEw Green: Dublin Bike availability Purple dot: Bus in congesHon Blue: Noise Purple bar: PolluHon Red: AmeniHes Yellow: Cameras

LD4SC-‐15

Seman&c Processing of Urban Data

h_ps://www.youtube.com/watch?v=lrUHet5awzw&feature=youtu.be

h_p://50.97.192.242:8080/Dali/

LD4SC-‐15

Cogni&ve Explana&on of Flight Delays

LD4SC-‐15

LD4SC-‐15

Conclusion

Ci&es are characterized by:

•  Big Data •  Complex Systems

Ci&es want to be Smarter: •  More efficient •  More reliable •  More secure •  More open

Ci&es can benefits from

•  REAL World data: Open Data to get Smarter •  Advances in AI

AI already helped a lot!! ... and should even contribute further •  Op&miza&on, coordina&on …

•  Cheaper •  Faster •  More integrated •  More ci&zen-‐centric

•  More a_rac&ve •  More Intelligent •  Sustainable •  Be_er city planning

•  Integrated Problems •  Scalability Challenges

LD4SC-‐15

Future Work Analy&cs and Reasoning: Scalability from One City to Another One

Picture of different ci&es

LD4SC-‐15

Future Work Applica&on: From Ci&es to mini-‐Ci&es

Airport

Supply-‐Chain Building

Stadium

Warehouse

LD4SC-‐15

Future Work More Mul&-‐disciplinary: AI (Planning, KRR, ML, …), Database, Mathema&cs …

More Science Integra&on

LD4SC-‐15

References (1) •  Freddy Lecue, Jeff Z.Pan. Consistent Knowledge Discovery from Evolving Ontologies. AAAI 2015. Aus&n, Texas, USA, January, 2015.

•  Anika Schumann, Freddy Lecue. Minimizing User Involvement for Accurate Ontology Matching Problems. AAAI 2015. Aus&n, Texas, USA, January, 2015.

•  Freddy Lecue, Simone Tallevi-‐Diotallevi, Jer Hayes, Robert Tucker, Veli Bicer, Marco Sbodio, Pierpaolo Tommasi. Smart traffic analy&cs in the seman&c web with STAR-‐CITY: Scenarios, system and lessons learned in Dublin City. In Jour-‐ nal of Web Seman&cs. Web Seman&cs: Science, Services and Agents on the World Wide Web (JWS). Vol XX (2014).

•  Spyros Kotoulas, Vanessa Lopez, Raymond Lloyd, Marco Luca Sbodio, Freddy Lecue, MarHn Stephenson, Elizabeth M. Daly, Veli Bicer, Aris Gkoulalas-‐Divanis, Giusy Di Lorenzo, Anika Schumann, Pol Mac Aonghusa. SPUD -‐ Seman&c Pro-‐ cessing of Urban Data. In Journal of Web Seman&cs. Web Seman&cs: Science, Services and Agents on the World Wide Web (JWS), Vol 24 (2014).

•  Freddy Lecue, Robert Tucker, Simone Tallevi-‐Diotallevi, Rahul Nair, Yiannis Gkoufas, Giuseppe Liguori, Mauro Borioni, Alexandre Rademaker and Luciano Barbosa. Seman&c Traffic Diagnosis with STAR-‐CITY: Architecture and Lessons Learned from Deployment in Dublin, Bologna, Miami and Rio, ISWC 2014, Riva, Italy, October 2014 (Best In Use Paper Award)

•  Joern Ploennigs, Anika Schumann, Freddy Lecue. AdapHng SemanHc Sensor Networks for Smart Building Diagnosis. In Proceedings of the 13th Interna&onal Seman&c Web Conference (ISWC 2014), pages 308-‐323, October 19-‐23, 2014, Riva del Garda, Italy. Springer 2014 ISBN 978-‐3-‐319-‐11914-‐4. (Best In Use Paper Award Nominee)

•  Jiewen Wu, Freddy Lecue. Towards Consistency Checking over Evolving Ontologies. Proceedings of the 23rd ACM Conference on Informa&on and Knowledge Management, CIKM 2014, Shanghai, China, November 3-‐7, 2014.

•  Anika Schumann, Freddy Lecue, Joern Ploennigs. Exploi&ng the Seman&c Web for Systems Diagnosis. In Proceedings of the Twenty-‐First European Conference on Ar& cial Intelligence (ECAI 2014), pages ??-‐??, August 18-‐22, 2014, Prague, Czech Republic. IOS Press 2014.

LD4SC-‐15

References (2) •  Joern Ploennigs, Anika Schumann, Freddy Lecue. Extending Seman&c Sensor Networks for Automa&cally Tackling Smart Building Problems. In Proceedings of the Twenty-‐First European Conference on Ar& cial Intelligence (ECAI 2014), pages ??-‐??, August 18-‐22, 2014, Prague, Czech Republic. IOS Press 2014.

•  Freddy Lécué: Towards Scalable Explora&on of Diagnoses in an Ontology Stream. AAAI 2014: 87-‐93 •  Freddy Lécué, Robert Tucker, Veli Bicer, Pierpaolo Tommasi, Simone Tallevi-‐Diotallevi, Marco Luca Sbodio: Predic&ng Severity of Road Traffic Conges&on Using Seman&c Web Technologies. ESWC 2014: 611-‐627 (Best In Use Paper Award)

•  Freddy Lécué, Simone Tallevi-‐Diotallevi, Jer Hayes, Robert Tucker, Veli Bicer, Marco Luca Sbodio, Pierpaolo Tommasi: STAR-‐CITY: seman&c traffic analy&cs and reasoning for CITY. IUI 2014: 179-‐188

•  Simone Tallevi-‐Diotallevi, Spyros Kotoulas, Luca Foschini, Freddy Lecue, Antonio Corradi. Real-‐Time Urban Monitoring in Dublin Using Seman&c and Stream Technologies. In Proceedings of the 12th Interna&onal Seman&c Web Confer-‐ ence (ISWC 2013), pages 178-‐194, October 21-‐25, 2013, Sydney, NSW, Australia. Springer 2013 Lecture Notes in Computer Science ISBN 978-‐3-‐642-‐41337-‐7.

•  Freddy Lecue, Jeff Z, Pan. Predic&ng Knowledge in an Ontology Stream. In Proceedings of the Interna&onal Joint Conference on Ar&ficial Intelligence (IJCAI 2013)

•  Elizabeth M. Daly, Freddy Lecue, Veli Bicer. Westland Row Why So Slow? Fusing Social Media and Linked Data Sources for Understanding Real-‐Time Traffic Condi&ons. In Proceedings of the ACM 2013 Interna&onal Conference on Intelligent User Interfaces (IUI 2013)

•  Freddy Lecue, Anika Schumann, Marco Luca Sbodio. Applying Seman&c Web Technologies for Diagnosing Road Traffic Conges&ons. In Proceedings of the 11th Interna&onal Seman&c Web Conference (ISWC 2012), Springer (Best In Use Paper Award Nominee)

•  Freddy Lecue: Diagnosing Changes in An Ontology Stream: A DL Reasoning Approach. In Proceedings of the Twenty-‐Sixth AAAI Conference on Ar&ficial Intelligence (AAAI 2012), July 22-‐26, 2012, Toronto, Ontario, Canada. AAAI Press 2012.

LD4SC-‐15

Ques&ons

Thank you!

Credits: Pol Mac Aonghusa, Luciano Barbosa, Veli Bicer, Antonio Corradi, Elizabeth Daly, Luca Foschini, Yiannis Gkoufas, Jer Hayes, Pascal Hitzler, Vanessa Lopez, , Raghava Mutharaju, Rahul Nair, Jeff Z. Pan, Joern Ploennigs, Alexandre Rademaker, Marco Luca Sbodio, Anika Schumann, Mar&n Stephenson, Simone Tallevi-‐Diotallevi, Pierpaolo Tommasi, Robert Tucker, Yuan Ren, Jiewen Wu

semantics for smarter cities

Science

cities source

food large cities

number of shrinking

number of expanding

habitat smart cities

large urban areas

greatest number of people

urban migration