Post on 22-Sep-2020
Big Data Breaking Barriers –First Steps on a Long Trail
Sven Schade (@innovatearth)
Institute for Environment and SustainabilityDigital Earth and Reference Data Unit
www.jrc.ec.europa.eu
Serving society
Stimulating innovation
Supporting legislation
36th ISRSE11 May 2015Berlin, Germany
Including contributions from:
Massimo Craglia, Davide De Marchi, Irene Eleta, Jacopo Grazzini,
Jiří Hradec, Alexander Kotsev, Frank Ostermann, Nicole Ostländer,
Francesco Pantisano, Elena Roglia, Cristina Sanchez, Sven Schade,
Spyridon Spyratos, Chrisa Tsinaraki, Lorenzino Vaccari, as well as
the two visiting scientists (Levente Juhász and Sergi Trilles)
Organisational Background -Digital Earth and Reference Data unit
• INSPIRE: Implementation, Maintenance and Evolution
• Open Data Policy of the JRC
• Complex data handling with geospatial informatics
2
Content
• Brief introduction to Big (Geospatial) Data
• Highlights from 10 completed case studies
• Discussion of 3 barriers
• Conclusions
3
This talk provides on overview of work that was largely carried out in
2014, with some reflections on a possible way ahead.
The content presents my personal view and does not necessarily reflect
the position of the European Commission.
[sourc
e:
jim
mie
joe.c
om
]
4Computer science & system engineering
[sourc
e:
big
data
landscape.c
om
]
[sourc
e:
NBD
RA, N
IST]
5
• Volume: remote sensing (images or point clouds in 2/3D), or intense modelling
(e.g. immediate and medium range weather forecasting and climate modelling)
• Velocity: large single volumes or continuous inputs - of the same type but from
massive amounts of sources (e.g. in the context of the IoT)
• Variety: given any place on earth (or elsewhere), we already
today receive spatially-related data sets and streams for
multiple sources. These grow and accumulate over time.
• Veracity: reference data and the differentiation between “authoritative” sources
and user-contributed contend discussed in the geospatial and
statistics communities
[sourc
e:
wik
imedia
.org
]
Geospatial information science
[sourc
e:
M. Cra
glia]
5
• Volume: remote sensing (images or point clouds in 2/3D), or intense modelling
(e.g. immediate and medium range weather forecasting and climate modelling)
• Velocity: large single volumes or continuous inputs - of the same type but from
massive amounts of sources (e.g. in the context of the IoT)
• Variety: given any place on earth (or elsewhere), we already
today receive spatially-related data sets and streams for
multiple sources. These grow and accumulate over time.
• Veracity: reference data and the differentiation between “authoritative” sources
and user-contributed contend discussed in the geospatial and
statistics communities
[sourc
e:
wik
imedia
.org
]
Geospatial information science
[sourc
e:
M. Cra
glia]
63D platform for geospatial data handling
3D browser based viewer
EU3D advanced desktop application
Both building on of the “Core003” data set, a Very High Resolution (VHR) optical coverage over the member and cooperating countries of the European Environment Agency (EEA) that was generated from SPOT-5 data through multi-spectral 2.5 meters resolution data ortho-rectified with a geo-location accuracy of less than 5 meters Root Mean Square Error (RMSE)
[source:D. De Marchi]
[source: D. De Marchi]
7Investigating usage potential of social media platforms
Using social network analysis to sense social behaviour
Using social media platforms to complement authoritative vector data
Using new database technologies to store and query social media data
[source:S. Spyratos et al.]
[source: I. Eleta, J. Grazzini and E. Roglia]
[source: L. Juhász]
8Sensing technologies and the Internet of Things
Service-enabled sensing platform for the environment
Real-time event detection from sensor networks
[source:S. Trilles et al.]
[source: A. Kotsev, F. Pantisano et al.]
9Handling complex data integration
New modes for multi-sensory integration
Visualisations of complex metadata
Model transparency
[source: J. Hradec]
[source: N. Ostländer]
[source:F. Ostermann and S. Schade]
10
Do not expect (only) one (Digital Earth) platform!
Establish data flows between existing systems and enable the exchange of lessons learned
Technical barriers
• Rich choice of implementations available
• Implications from technical choices and re-use
11Discussion
[sourc
e:
pix
abay.c
om
]
Enable knowledge transfers across the sciences
Do not (only) work together!
• Potential useful data does not reside in
well-know communities any more
• Desired to use data and tools from
other sciences and vice versa
12
Semantic barriers
[sourc
e:
pix
abay.c
om
]
Discussion
13
Organisational barriers
• Beneficiaries of applications do not only
reside in (data) science
• Adding value by including “foreigners”
and opening new “markets”
Increase efforts on social and behavioral aspects
Do not (only) address the technocratic dimension! [s
ourc
e:
pix
abay.c
om
]
Discussion
• These are only a few examples and findings
• Many case studies have been developed across the globe
• Many of us took their first steps on the long trail of successful and useful
knowledge extraction from (small and big) data
14Conclusion
[source: wikimedia.org]
• These are only a few examples and findings
• Many case studies have been developed across the globe
• Many of us took their first steps on the long trail of successful and useful
knowledge extraction from (small and big) data
We are currently investigating the use of RM-ODP to integrate the findings of any number of case studies.It seems promising to use this standard methodology to describe information systems that is already widely used in the geospatial information domain to develop a high-level description of Digital Earth platforms and provide guidance for case specific re-use.
14
[source: pixabay.com]
Conclusion
[source: wikimedia.org]
• These are only a few examples and findings
• Many case studies have been developed across the globe
• Many of us took their first steps on the long trail of successful and useful
knowledge extraction from (small and big) data
Final slide, reallyThank you!
[source: wikimedia.org]
We are currently investigating the use of RM-ODP to integrate the findings of any number of case studies.It seems promising to use this standard methodology to describe information systems that is already widely used in the geospatial information domain to develop a high-level description of Digital Earth platforms and provide guidance for case specific re-use.
[source: pixabay.com]