Webinar - INSPIRE 2020 Virtual Conference
----
How Spatial Data Infrastructures (SDIs) can evolve into Data Ecosystems?
This is the main question that the ongoing study addressing “Data ecosystems for geospatial data - Evolution of Spatial Data Infrastructures” (JRC/IPR/2019/MVP/2781) is addressing. It is performed by the Luxembourg Institute of Science and Technology (LIST) in close collaboration with the Joint Research Centre of the European Commission.
The purpose of this study is to identify and analyse a set of successful data ecosystems and to address recommendations in support of the implementation of data-driven innovation in line with the recently published European Strategy for Data. It investigates factors such as relevant actors, their responsibilities and data value chains, emerging data sources (e.g. the Internet of Things) and technical/architectural approaches (e.g. digital platforms, mobile-by-default, Application Programming Interfaces). It also addresses the interoperability between data ecosystems in different sectors and/or different countries and crosscutting requirements for geospatial data.
This session is intended to share with the audience the study approach, methodological approach and first identified Data Ecosystems, and to learn from their experiences with Data Ecosystems: emergence, barriers, opportunities, sustainability, interoperability between ecosystems, etc.
AGENDA
14.00 - Welcome, Introduction to the context of the study (JRC)
14.10 - Study approach and methodological framework (LIST)
14.20 - Identified data ecosystems and selection criteria (LIST)
14.25 - Illustration of data ecosystem analysis (LIST)
Ghislain Delabie, Simon Saint-Georges, Urban Rennes Data Interface
Sean Wiid, UP42
Charles Moszkowicz, ENEO • Interactive session (All, 20)
Next activities, Goodbye.
2. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/27812
Data ecosystems for geospatial data -
JRC/IPR/2019/MVP/2781
3. Mute your mic!
To mute and unmute,
click the microphone icon
next to your name or at
the bottom of the screen.
Turn off video
Share your webcam
video only when you are
talking. To do this, click
video icon next to your
name.
Ask a question
Use “raise hand”
functionality to ask a
question. Click the hand
icon next to your name in
the participant list. If this
is not available write
‘hand’ in the chat.
Welcome and some hints for participants
4. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
“Data ecosystems for geospatial data - Evolution of Spatial Data
Infrastructures”
• Ongoing study, started in January 2020
• Performed by the Luxembourg Institute of Science and Technology
(LIST) in close collaboration with Joint Research Centre of European
Commission.
• Identify and analyse a set of successful data ecosystems and to
address recommendations in support of the implementation of data-
driven innovation in line with the recently published European strategy
for data.
Sharing the methodological approach undertaken;
Presenting identified data ecosystems
Call for:
• Ecosystem use cases
• Identifying & reaching ecosystems' experts
Rationale of the session
4
5. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander
Kotsev, JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
5
6. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data”
- Alexander Kotsev, JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
6
7. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Alexander Kotsev, JRC
Landscape of the study
8. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
INSPIRE - STATE OF PLAY
8
• The INSPIRE Directive is 13 years old (now formally a teenager)
• Deadlines for full implementation are approaching
• What has changed in the past few years, and what are our
outstanding challenges?
1. Technological perspective
2. Organisational perspective
3. Political perspective
9. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
We have come a long way!
1. TECHNOLOGICAL PERSPECTIVE
9
Technology in 2007
10. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
API4INSPIRE
260+ million spatio-temporal, pan-European
measurements
• standard-based access through OGC
SensorThings API
• data provided in a public cloud through
virtualization
• harvested air quality data from multiple sources:
• national INSPIRE download services
• European Environment Agency
1. TECHNOLOGICAL PERSPECTIVE
10
http://www.datacove.eu/ad-hoc-air-quality/
Technology in 2020
11. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Heterogeneous data sources
○ Citizen data / personal data
○ Remote sensing (e.g. Copernicus)
○ IoT
2. New technologies
○ Data handling at the edge/fog
○ Virtualisation and cloud computing
○ From data collection to data connection (APIs)
3. New standards
○ Embracing web best practices
○ Following an agile and inclusive approach
■ SensorThings API
■ OGC API - Features
1. TECHNOLOGICAL PERSPECTIVE
11
12. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1) Who does what in the European context?
○ Distributed system
■ 7000+ data providers (tip of an iceberg)
■ Federated governance
■ Excellence on subnational level
○ Emerging agile approaches at multiple levels
■ Hackathons, code sprints
■ wikifikation / Git repositories
2) Resources for SDI and sustainability of infrastructures
○ Many developments are based on projects
■ Projects do end
3) How to modernise/update existing infrastructures
■ Technologies changes are fast
■ Procurement and organisational changes are not
4) “Follow the user”
○ Sure, but how?
2. ORGANISATIONAL PERSPECTIVE
12
13. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1) Europe Fit for the Digital age
○ European Strategy for Data
• Establish a pan-European single market for data
• Regulatory sandboxing
• Emphasis on the benefits of different actors
• Sector-specific data spaces
○ White paper on AI
• Extensive reuse of available data
○ Open Data Directive
• High-value datasets (exposed through APIs)
2) European Green Deal
o GreenData4all initiative
• Modernising INSPIRE
o Destination Earth
3. POLITICAL PERSPECTIVE
13
14. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• The INSPIRE community is healthy (and growing)
• There are favourable conditions for data-driven
innovation in Europe!
Q: Is spatial still special?
A. Yes, everything that happens happens
somewhere. Location information is fundamental
for an increasing number of use cases.
B. No, SDI developments should be merged into
mainstream ICT.
● We should
○ Avoid that SDIs become a big silo
○ Focus on sustainability & scalability
○ Learn from existing ecosystems
The context for the evolution of SDIs
14
FROM SDIs TO DATA ECOSYSTEMS?
15. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
15
16. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Luxembourg Institute of Science and
Technology (LIST)
Prune GAUTIER
Sébastien MARTIN
Slim TURKI
17. Research and Technology Organization (RTO)
Develops innovative and competitive solutions in
response to the key needs of Luxembourgish and
European economies.
• Employees: ~600 | Budget: EUR 66 millions
• Activities:
• Fundamental and applied scientific research,
development of knowledge and competences;
• Experimental development, incubation and transfer of
new technologies, competences, products and services;
• Scientific support to the policies of the Luxembourgish
government, businesses and society in general;
• Doctoral and post-doctoral training, in partnership with
universities.
LUXEMBOURG INSTITUTE OF SCIENCE AND
TECHNOLOGY
17
Interdisciplinary portfolios
• Smart cities
• Spatial sector
• Industry 4.0
• FinTech and RegTech
Fields of activity
• Digital innovation
• Ecological innovation
• Materials innovation
Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
18. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Objectives of the study
18
Investigate how Spatial Data Infrastructures (SDIs) can evolve into
data ecosystems to support the goals of digital government in Europe.
• Take into account factors such as relevant actors, their responsibilities and data value
chains, emerging data sources (e.g. the Internet of Things) and technical/architectural
approaches (e.g. digital platforms, mobile-by-default, Application Programming Interfaces).
Address the interoperability between data ecosystems for different
sectors and/or different countries and cross-cutting requirements for
geospatial data.
Provide an input into the discussion on the future evolution of INSPIRE
after the conclusion of the current implementation programme in 2020.
19. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Identify existing Data Ecosystems and case studies for
interoperability between such Data Ecosystems
Analyse / compare characteristics / requirements of Data
Ecosystems and their interoperability
Analyse in depth a subset of Data Ecosystems
Develop recommendations for setting up Data Ecosystems and to
enable interoperability between them
Work plan
19
> Desk research
> Academic literature
> Reports (official reports, projects reports, etc.)
> Companies' documentation
> Qualitative data (interviews)
20. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Technology
Policies
Standards, and
Human resources
Necessary to
o Acquire
o Process
o Store
o Distribute, and
o Improve utilization of Geospatial data
• Linear process in which data is published
and made discoverable and usable
• No feedback loop between users and
providers - critical to the potential
sustainability and evolution of data
ecosystems.
SPATIAL DATA INFRASTRUCTURES (SDIs)
Definition
20
21. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
DATA ECOSYSTEMS
21
EcosystemDataEcosystem
Evolve and adapt through a cycle of
data creation and sharing, data analytics,
and value creation in the form of new products,
services, or knowledge, which, when used,
produce new data feeding back into
the ecosystem.
Complex socio-technical system of People,
Organizations, Technology, Policies and Data In
specific Area/Domain, that Interact with one
another and their surrounding environment to
achieve a specific Purpose
Data ecosystem =
Ecosystem analysed with a
strong focus on data issues
22. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• How Ecosystem Thinking may contribute to identify and
trigger the uptake of SDIs?
• What is the Position and Role of SDIs in Self-sustainable Data
Ecosystems?
• Addressing interoperability between Data Ecosystems
WHAT?
22
23. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, Simon Saint Georges
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
23
24. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• Looking for cases studies illustrating shifts from Data Infrastructures to
Self-sustainable Data Ecosystems
• With potential to enrich recommendations
• Diversity of the case studies/use cases is important.
• Regional/City, if possible with Public, Private sectors and Citizen
involvement.
• Thematic: Mobility, Agriculture, Insurance, etc.
• Generic, involving user data and feedback, such as recommendation in
tourism
• Community oriented, such open science networks
• Place / role of spatial data
• Where expert(s), documentation or standards are available and
accessible
DATA ECOSYSTEMS IDENTIFICATION
24
25. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Evaluation
DATA ECOSYSTEMS IDENTIFICATION
25
26. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
DATA ECOSYSTEMS IDENTIFICATION
26
Rennes Urban
Data Interface
Tracking Technologies
for Supply Chain
Hotel reviews
Tripadvisor, Yelp,
etc.
VOD Entertainment
(Netflix, Disney, etc.)
Vehicles and planes
fleets predictive
maintenance
Circular
economy
Pandemic
Data
Weather
Forecast
Digital patient record
(e-health)
B2C
e-commerce
Data
Marketplace
Smart
Agriculture
ML services on
space imagery
Pan-European Invasive
Alien Species Monitoring
and Reporting
Crowdsourced traffic
information (Waze)
Energy efficiency of
buildings
Location aware
dating services
National wide
data ecosystem
27. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• Sought expertise
• General overview of the Data Ecosystem
• Specific (actual ecosystem participants speaking from their perspective)
• Kinds of experts
• Data owners (public and private);
• Data re-users (public and private)
• Platform actors
• Academic
• (End-users)
• Available for some online/onsite interviews
Looking for Experts
27
28. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Q1: Do you know about Self-sustainable Data Ecosystems?
• With potential to enrich recommendations
• Where expert(s), documentation or standards are available and
accessible
Q2: Would you kindly recommend experts of already identified data
ecosystems?
Please share your suggestions
• Slim TURKI, LIST, slim.turki@list.lu
• Alexander KOTSEV, JRC, alexander.kotsev@ec.europa.eu
• Using the Chat
IDENTIFICATION of DATA ECOSYSTEMS
28
Suggestions?
29. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework -
LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
29
30. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• How the ecosystem thinking may
contribute to identify and trigger the
uptake of spatial data infrastructures?
• (Oliveira & Loscio, 2018) : “Lack of well-
accepted definition of the term Data
Ecosystem”
• Ecosystem is a paradigm, to analyse a
network and to act on it.
• Ecosystem emergence (Thomas, 2015)
• Ecosystem health
• Increasing data-reuse beyond the
original purpose
• New modes of data creation
• Case of non-geospatial data
ECOSYSTEM THINKING
30
31. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
(Self-) sustainability:
• Well-balanced ecosystem
• Able to function and development without
direct government support
Combination of supportive factors:
• Value creation and business model
• Value distribution
FOCUS ON BUSINESS MODELS &
VALUE CREATION
31
Value Creation
Business
Models
32. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• Orchestrating an ecosystem means managing the
network relations.
• Originating from innovation networks (Gawer &
Cusumano, 2014).
• Keystone actor(s) / actor(s) leadership
• Orchestration concepts
• Ecosystem membership (size, diversity)
• Ecosystem structure (density, autonomy)
• Ecosystem position (centrality, status)
• Appropriability regime
• Knowledge mobility
• Ecosystem stability
• Kinds of orchestration
• Organizational orchestration
• Technical orchestration
• Standard / industry standard adoption
• Internal / external interoperability
• But intertwined
FOCUS ON ORCHESTRATION
32
Orchestration
33. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Modular
Analysis
Framework
A MODULAR ANALYSIS FRAMEWORK
33
1
2
3
> Desk research
> Academic literature
> Reports (official repor
ts, projects reports, etc.)
> Companies'
documentation
> Qualitative data
(interviews)
34. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
34
Ecosystem Summary
Provides an overall representation of the
components of the ecosystem.
1
Focuses on Data Ecosystems key aspects
• Goal / purposes
• Main actors, their exchanges and communication
• Legal context and governance
• Technology specific aspects
• Cost and revenues / benefits
• Faced Barriers and incentives
First level of assessment, inspired by the
Business Model Canvas from Alexander
OSTERWALDER
35. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
35
Ecosystem Dynamics
Represents the interactions
between stakeholders.
2
The graphical representation of this second layer of the
Framework finds its inspiration in network modelling tools.
While illustrating the resources exchanged between the
stakeholders, and their value, we can:
- follow the value creation,
- highlight the orchestration
- and evaluate the sustainability of the ecosystem ( balanced
counterparts missing, actor missing, value lost).
Goal centred, it gives a representation of
the actors commitment level and strength of interactions
36. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
36
Ecosystem Data flows
Focuses on the associated data
flows / Data life cycles
3
• Data flows as proxy of the maturity and health of an
ecosystem
• Data cycle: (Pollock, 2011) "infomediaries — intermediate
consumers of data such as builders of apps and data
wranglers — should also be publishers who share back
their cleaned / integrated / packaged data into the
ecosystem in a reusable way — these cleaned and
integrated datasets being, of course, often more valuable
than the original source.“
• Approach and graphical representation are derived from
product lifecycle model
37. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Q: Do you agree on the dimensions?
Q: Do you see missing dimensions?
A MODULAR ANALYSIS FRAMEWORK
37
38. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
38
Ecosystem Summary
Provides an overall
representation of the
components of the
ecosystem.
1
39. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
39
Ecosystem Dynamics
Represents the interactions
between stakeholders.
2
40. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
A MODULAR ANALYSIS FRAMEWORK
40
Ecosystem Data flows
Focuses on the associated data
flows.
3
41. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin,
ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
41
43. RUDI - Rennes Urban Data
Interface
An Open and Inclusive Metropolitan Data Ecosystem
44. Increasing interoperability, ease of access and
network effects among local/urban stakeholders
Urban Data Interface
Users
Citizens
Data producers
Projects
Increasing interoperability, ease of access and
network effects among local/urban stakeholders
Ecosystem
dynamics
Governance
Public debate
Open Data
RGPD policy
Public digital
services
Rennes pioneered Open Data in 2010.
Many Data of public interest are private
and could foster innovation at the city
scale
45. Innovators & service
providers
Increasing interoperability, ease of access and
network effects among local/urban stakeholders
Connecting innovators and citizens at the local level
Technical architecture
Rennes pioneered Open Data in 2010.
Many Data of public interest are private and
could foster innovation at the city scale
Citizens & users
Personal Data in a
secure space
RGPD compliance
Ecosystem management
Local Data
46. External partners &
Data Ecosystems
RUDI Data Ecosystems
Partner IS
Partner 2
Partner IS
Partner 1
Users
Portal + meta-catalog
RUDI helps local innovators, service providers and Data providers
to cooperate around Data to produce new services and broadcast
Data
RUDI provides connection between various Data ecosystems (e.g.
mobility, waste management) and enables new Data ecosystems in
new domains/areas of local interest
50. A developer platform and marketplace
to derive insights from geospatial data at
scale
50
What is UP42?
UP42 Overview
51. 51
Founded and 100% owned
by:
Incubated and mentored by:
May
2019
Company & Beta
Launch
First
Revenue
July
2019
Commercial
Launch
Sept
2019
Catalog
Search
Dec
2020
1 Year
Anniversary
May
2020
45 people from 21
countries based in
Berlin
UP42 Overview
52. What problem does UP42
solve?
It’s hard to integrate data
sources.
Data Sources
It’s hard to develop processing
algorithms.
Processing algorithms
It’s hard to set up a compute
infrastructure.
Infrastructure
UP42 Overview
52
54. Processing
blocks3rd party and/or
custom
Workflows
Combination of blocks
acting as a template for
jobs. Jobs
Run separately and in parallel at
scale.
Results
The results of the jobs are
valuable, domain specific
insights.
3rd party and/or custom
Data blocks
How does it work?
UP42 Overview
54
55. UP42 is paradigm shift in the industry
UP42 Overview
55
Workflow Engine & Scalable Infrastructure
Process data at scale with confidence. Our platform ensures data and
algorithm compatibility and our compute infrastructure scales up and
down automatically.
Powerful Platform APIs and SDKs
Developers can access the full power and scalability of the UP42
platform and integrate directly into products or data analytics
operations. The open source Python SDK includes many helper
functions and supports Jupyter Notebooks integration.
Simple UI for Discovery & Rapid Prototyping
Anyone can search and preview data across multiple providers, order
archive and tasked imagery, build data processing workflows and
extract insights from the data at scale without needing to write a single
line of code.
Open Marketplace of Geospatial Data & Algorithms
Users can discover and immediately access data and analytics from
the industry's leading geospatial companies and pay only for what they
use. Transparent, simple pricing. Partners accrue revenue share on
every use of their block, no matter how small the individual transaction.
1
2
3
4
56. Our marketplace has 25+ data blocks so
far
UP42 Overview
Example Commercial data sources Example Open data sources
56
Sample data
57. Our marketplace has 50+ algorithms so far
UP42 Overview
Data preparation & pre-
processing examples
Indices, bandmath &
statistics examples
AI/ML based object
detection & classification
examples
57
58. Easily click together workflows and
run analytics at scale using our
APIs
UP42 Overview
Data
349 Sentinel scenes
in parallel
Processing
5.9 TB of processed
tiled SAR data
Infrastructure
~870 VCPU Cores
~7TB of Memory
Performance
~60 minutes
end to end
58
60. 60
1. Building eco-systems means
collaboration
3 Lessons from Building a Geospatial Marketplace
Revenue share agreements
and/or “pay for what you use”
are often new business models
for data owners.
Protecting existing business
often conflicts with accepting
these new ways of working.
New business models
Technical requirements to
onboard onto a marketplace can
vary.
Data owners are wary of
spending resources on
deploying to a new marketplace
without knowing the ROI upfront
or taking an upfront fee.
Integration costs
Data owners often want to be
“exclusive” on marketplaces.
However, this is against a key
principle and success factor of
open marketplaces and
ecosystems: Offer alternatives
and let the customer decide
Co-existing with
competitors
Getting data owners to take part in a marketplace often means taking them out of their comfort zones
and challenging the established rules of business. This can take time and should not be
underestimated.
61. 61
2. Data owners requirements are central
3 Lessons from Building a Geospatial Marketplace
Data owners need to have full
control over how their products
are presented in the
marketplace.
This should include all
metadata, technical data and
use case descriptions.
Metadata & Marketing
Data owners need to be able to
establish or at least have input
into the end-user price on the
marketplace.
Otherwise, the marketplace
could become a channel that
undercuts prices and erodes
value over time.
End User Pricing
Data owners need to be able to
propagate their own EULAs to
the end user.
This is necessary to ensure that
other channels, regulatory
requirements, exclusive reseller
agreements etc are all
respected downstream.
End User License
A marketplace can only be successful in the long term if it gains the trust and confidence of its
suppliers as a safe place to do business. This means that a marketplace must put a lot of control in
the hands of data owners.
62. 62
3. Data platforms still need to mature
3 Lessons from Building a Geospatial Marketplace
For some data sources, e.g.
Copernicus data, many different
organisations have created
partial copies of the data archive
This leads to duplication of large
amounts of data with none
providing full access to the
whole archive at sufficient scale.
Fragmented & Incomplete
Not all data sources have
adequate metadata or support
standards for metadata
catalogs.
Resolving this makes it much
easier to “plug” into existing
ecosystems & marketplaces.
Catalog & Search
Not all data sources support
search, order and delivery via
API.
APIs are the glue that hold
ecosystems together, and we
believe strongly that any data
marketplace and data provider
should treat developers as their
main customer persona.
APIs
Data is still too fragmented, delivered using too many different formats and in many cases without a
modern search and data delivery APIs
64. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• Explorama : a treasure hunt app to discover Nature
• Spipoll : collaborative science project to collect data on insects
• Foxtrot : Green itinerary generator for
Orleans city area
64
Digital agency for Nature discovery
ENEO
65. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Maximize your contact with Nature
Foxtrot algorithm - ENEO
65
66. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• OpenStreetMap Data
• footpaths, cyclepaths, roads,
reduce mobility access…
→ Wide spread
→ Community based
→ Standardized
→ Documented
66
2 main sources of data
Foxtrot algorithm - ENEO
• Orleans Metropole Data
• trees, parks, known walks, heatmap,
noisemap…
→ Added value
→ Many partners involved
→ OpenDataSoft based
→ Data quality not quite there yet
67. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Foxtrot algorithm - ENEO
67
Foxtrot schema
Aggregate data → run algorithm → outputs
68. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Process steps
Foxtrot algorithm - ENEO
68
• Extract roads network
• Apply weights
Parks
Water areas
Trees
Noise
Bench
Playground
Scrubs
69. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/278169
Shortest path « Greenest » path
70. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Technical and format issues for spatial data
• coordinates: lon-lat lat-lon
• projections: meters degrees
• format: geojson != json
• warning : point != polygons != linestrings
Availability issues
• new datasets has the project is running.
• some datasets aimed at professional but lack of relevant information for wide
audience
Advice: Docker is fire !
• Build complex and stable architecture quickly
• Automate data download and process replicability, ease update
70
Issues
Foxtrot algorithm - ENEO
71. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
For you
• Open-source
• Made for replicability in other areas
• Open to external request through REST API
For us
• Generate itinerary data
• Learn about users habits, profiles
• Extend usage to involve new partners like tourism, local farmers...
71
Next steps
Foxtrot algorithm - ENEO
72. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
72
73. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Questions?
Discussion - Interactive session
73
74. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
1. Landscape of the study « Data Ecosystems for Geospatial Data - Alexander Kotsev,
JRC
2. Study objectives and workplan - LIST
3. Data Ecosystems Identification and Selection - LIST
4. Ecosystem Thinking & Modular methodological framework - LIST
5. Illustration of data ecosystem analysis with experts
Rennes Urban Data Interface - Ghislain Delabie, OuiShare
Machine Learning for Geospatial Data - Sean Wiid, UP42
Data cycles, feedback from the field - Jean-Charles Simonin, ENEO
6. Interactive session
7. Conclusion & Next activities
AGENDA
74
75. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
Conclusion & call for participation
75
Next steps:
• Experts interviews, desk analysis, documentation
review
• Selection and in analysis depth analysis of 5
ecosystems
• Recommendations for setting up self-sustainable data
ecosystems
• Stakeholders workshop to present and challenge the
findings (Fall 2020)
76. Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781Data ecosystems for geospatial data - JRC/IPR/2019/MVP/2781
• Joint Research Centre:
• Alexander KOTSEV, alexander.kotsev@ec.europa.eu
• Luxembourg Institute of Science and Technology
• Slim TURKI, slim.turki@list.lu
Contacts
76