"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
Data-in-the-Cloud City
1. Data-in-the-Cloud
City
Proactive Analysis of Digital
Information about the city ! !
Gonzalo A. ARANDA-CORRAL Alejandro BLANCO-ESCUDERO
Universidad de Huelva Yaco Sistemas
Department of Information Technology alejandro.b.e@gmail.com
gonzalo.aranda@dti.uhu.es
Joaquín BORREGO-DÍAZ Manuel GOMAR-ACOSTA
Universidad de Sevilla Elelog S.L.
Dept. of Computer Science and AI mangomaco@gmail.com
jborrego@us.es
2. Index
Motivation & Goals
Data in the WWW and associated services
Simulating extreme dynamics
Multiagent Arch
Results
Conclusions and Future Work
3. Motivation (I): Context
eCompleXcity
Emergent concepts in complex systems. Applications
to Urban environments and Cultural Complexity
Excellence Project. Junta de Andalucía. Spain
4. Motivation (II): Digital Information
So
c ia 0
lM 2.
Marketing
ed eb
ia W
Heterogeneity
Different nature
Architecture Urbanism
Goal-driven
Different information
flows
ns Lo
tio
Media Art
c
¿Reusable? ic
a
at
se ion B
rv
un ice ase
s d
m
com
le
Te
5. Motivation (III)
Urban dynamics
simulated from WWW
data
he
nt r
Multiagent Systems a i fo
at d
D
(MAS) for simulating C lou an
Complex Behaviour from U rb cs?
mining WWW information na mi
Dy
Limits of MAS simulation
from Data about cities
6. City as a Complex System
Different views:
Data city
Social Network city
City as a ground of
cyberinfrastructure
Local Interaction versus
Global interaction !
7. Emergent Research Line
Collect and process
data for
new applications,
services, and
planning
Analysis of urban
behaviour
Open Data initiatives
facilitate R&D initiatives
8. Some questions...
How are WWW data about a
city?
What about the quality?
Are they useful?
Can they be improved?
9. pre-Digital Cities versus
Smart Cities
pre-Digital City:
¿Able to consume
data?
Smart City:
To produce and
consume its own data
10. Data Flows
I2U U2U
Op
ial ks
Da en oc or
S w
ta
N et
y
ilit
rab U
Info rban
pe
ero rm
I2I Int atic
s U2I
11. Data flows about cities in WWW
(I): Institutions to User (I2U)
Essential to understand
some urban process
(dynamics)
Historical data and
analisys
Main support of
Opendata.
12. Data flows about cities in
WWW (II): User to user
U2U (entre usuarios): P2P
Mobile devices and Social Web
Information quality.
13. Data flows about cities in
WWW (III): User to Institutions
U2I
Strong Growth
Web 2.0 & Urban informatics
14. Data flows about cities in WWW
(IV): Institutions to Institutions
Unavailable to users
Goverment (&
enterprises)
interoperability
Increasing
15. Different data sources for
MAS simulation
Extreme Urban dynamics Explore every WWW
information about both
Urban evolution under
the city and the event
exceptional
circunstances Data comsumption by
MAS
pre-Digital city: New
Orleans
Extreme dynamics:
Katrina hurricane (2005)
16. Why this event?
First, Katrina is one of the There exists a big amount
most destructive of data source and Web
hurricane suffered by a services associated (or
developed country, USA consumable by)
Geographic Information
The extent of damage
Systems with public
invites for a macroscopic
access
analysis of the incident
17. Bounding the scope
In order to evaluate the In some cases a
quality, accessibility and reparation of defficent
usefulness of I2U data is necessary
It considers only I2U Mainly, data from global
accessible by WWW, information systems (or
Internet or deep Internet U.S. Agencies)
( that is, accessible via
more specific data may
search forms)
limit the reusability.
18. Why MAS?
MAS based simulation
methodology allows to
estimate how affect data
quality to each module of I2U main flow for this
the system: simulation
Statistical results from
surveys useful for agents-
citizens behaviour.
19. I2U about New Orleans,
Katrina and its effects (I)
U.S. Geological Survey
(http://www.usgs.gov/)
National Elevation Dataset
(http://ned.usgs.gov/).
Precisions ~ 3 meters
Open Street Maps (OSM,
http://
!
www.openstreetmap.org/),
22. Environment agents
Information about terrain
Discretized in hexagons
Update water information
(by WaterAgent request)
Citizen Agents ask
information to
environment agents
23. Water agents (I)
Potential energy: Reactive
agent
Direction
Speed
Unaffordable information
River is initial agent state
24. Water agents (II)
Future;
Buildings geometry ! !
http://sketchup.google.com/
Complemented by OSM.
25. Citizen agents
Papers about social Also design groups of
behaviour in critical agents
situations
Based on published
Fundamental to citizen information
agents design
MAS level
Patterns of behaviour
Evacuation paths
from the surveys of
survivors Group behaviour in
panic situations
Prevent riskies situations
26. Visualization
Based on OSM and
Google Maps/Earth
Some extra data:
Disaster scope
Survivors / zone
etc...
http://www.youtube.com/watch?v=pTKhrpl9jZc
28. Conclusions
I2U information flows is used, in this work, to simulate
urban phenomena
Simulation needs digital information from cities and its
own feedback
Previous and exhaustive information analysis and
clasification its fundamental to start any kind of urban
cloud computing
29. Future Work
Use Complex Systems
methodologies to analyse and
to compare events and
simulations
To detect emergent
phenomena in digital cities by
means of simulation and Data
mining
30. Data-in-the-Cloud
City
Proactive Analysis of Digital
Information about the city ! !
Thanks for your attention