Big data is expected to have a profound economic and societal impact in mobility and logistics. Examples include 500 billion USD in value worldwide in the form of time and fuel savings, and savings of 380 megatons CO2. With freight transport activities projected to increase by 40% in 2030, transforming the current mobility and logistics processes to become significantly more efficient, will have a profound impact. A 10% efficiency improvement may lead to EU cost savings of 100 billion EUR. This keynote will highlight the key value dimensions for big data in mobility and logistics. The talk will present examples from the EU Horizon 2020 lighthouse project TransformingTransport, demonstrating the transformations that big data can bring to the mobility and logistics sector. TransformingTransport addresses 13 pilots in seven highly relevant pilot domains within mobility and transport that will benefit from big data solutions and the increased availability of data. The talk will close with an outlook on barriers and future opportunities.
Scaling API-first – The story of a global engineering organization
Big Data Value in Mobility and Logistics
1. Big Data Value in
Mobility and Logistics
Andreas Metzger
(TT Technical Coordinator)
2. Agenda
1. TT and the Big Data Value Ecosystem
2. TT Methodology
3. Transport Innovation via Big Data
4. Future Opportunities and Barriers
BMDA 2018, Vienna 2
3. About TT
• EU Horizon 2020 Big Data Value PPP Large Scale Pilot Action
• Goal: demonstrate transformations big data has on mobility and logistics
• 46 members - 18.7 MEUR budget - 30 months duration
BMDA 2018, Vienna 3
4. About TT
13 pilots in 7 domains
BMDA 2018, Vienna 4
Smart Highways Smart Airport
Turnaround
Ports as Intelligent
Logistics Hubs
Proactive Rail
Infrastructures
Sustainable Connected
Vehicles
Integrated Urban
Mobility
Dynamic Supply
Networks
5. Big Data Value Ecosystem
Big Data Value means…
• Achieving socio-economic impact with Big Data
• Increased efficiency, higher customer satisfaction, new business models, …
ETP4HPCEOSC ECSO
AIOTI
5G PPP
EFFRA
BMDA 2018, Vienna 5
6. Big Data Value Ecosystem
Big Data Value Association (BDVA)
• Industry-led; 55% industry
• > 180 members from 28 different EU countries
• Annual Research & Innovation Agendas
Technical Priorities
• Data Management
• Data Processing Architectures
• Data Analytics
• Data Visualisation and User Interaction
• Data Protection
• Engineering & DevOps for Big Data
• Big Data Standardisation
Non-technical Priorities
• Skills development
• Ecosystems and Business Models
• Policy and Regulation
• Social perceptions and societal implication
BMDA 2018, Vienna 6
7. Big Data Value Ecosystem
Big Data Value Public-Private Partnership (BDV PPP)
• European Commission (public side) + BDVA (private side)
• Implemented through calls for actions / projects under Horizon 2020
• Current work programme (WP2018-2020)
– https://ec.europa.eu/programmes/horizon2020/en/h2020-section/information-and-
communication-technologies
• Types of projects:
BMDA 2018, Vienna 7
Data Platforms
(Personal/Industrial)
WP Topic:
ICT-13a/b-WP2018-2020
Lighthouse Projects
(Large-scale Pilots / Test-beds)
WP Topic:
ICT-11/14-WP2018-2020
Technical Projects
WP Topic:
ICT-12@WP2018-2020
Collaboration & Support Actions
WP Topic: ICT-13c@WP2018-2020
8. Agenda
1. TT and the Big Data Value Ecosystem
2. TT Methodology
3. Transport Innovation via Big Data
4. Future Opportunities and Barriers
BMDA 2018, Vienna 8
9. TT Methodology
Rationale
• “No free lunch”[1]
– Each data set, domain, use case is different
– Using a single data analytics solution will
most probably not work
• Thus: For each of the 13 Pilots
– Dedicated data analytics solutions best suited for requirement and
datasets
– Dedicated infrastructures best linked to data sources
• Still: Reuse of do‘s/don‘ts, best practices, common
requirements, lessons learned, …
– Within pilots, across pilots, beyond project
[1] David Wolpert, William G. Macready:
No free lunch theorems for optimization. IEEE
Trans. Evolutionary Computation 1(1): 67-82
(1997)
BMDA 2018, Vienna 9
11. TT Methodology
3-Stage validation and scale-up
Stage Embedding Scale of Data
Technology
Validation
Problem understanding and
validation of key solution ideas
(Historic) data pinpointing
problems and opportunities
Large-scale
Experiments
Controlled environment (not
productive environment)
Large historic and real-time data,
possibly anonymized / simulated
In-situ (on site)
trials
Trials in the field, involving actual
end-users
Real-time, live production data
complementing historic data
BMDA 2018, Vienna 11
12. Agenda
1. TT and the Big Data Value Ecosystem
2. TT Methodology
3. Transport Innovation via Big Data
4. Future Opportunities and Barriers
BMDA 2018, Vienna 12
13. Transport Innovation via
Big Data
13
(Icon Source: DHL/DETECON)
Efficiency
Customer
Experience
Business
Models
Smart Highways ++ ++ o
Sustainable Connected Vehicles ++ ++ o
Proactive Rail Infrastructures ++ + o
Ports as Intelligent Logistics Hubs ++ + o
Smart Airport Turnaround ++ + +
Integrated Urban Mobility ++ ++ o
Dynamic Supply Networks + + +
New
Business
Models
Improved
Operational
Efficiency
Better
Customer
Experience
BMDA 2018, Vienna
Key Value Dimensions:
14. Transport Innovation via
Big Data
Data-driven decision making in retailing
@ Athens International Airport
14
Advanced big data
analytics solutions
(Indra INPLAN) to
anticipate
passenger flow and
preferences
Adapt marketing to
expected passenger
typology per time
slot
Use data insights to
exploit market
niches
BMDA 2018, Vienna
15. Transport Innovation via
Big Data
15BMDA 2018, Vienna
Advanced analytics
solutions (Indra
HORUS) for improved
traffic distribution
along road corridor
Better information
and decision tools for
road users
Real-time incident
warnings based on
novel sensor
technology
Improved driving and travel experience
@ CINTRA/Ferrovial-managed highways
16. Transport Innovation via
Big Data
16BMDA 2018, Vienna
Run-time
visualization of
operations to
increase terminal
productivity
Predictive analytics
to generate warnings
for proactive
transport
management
Enhanced decision
support for terminal
operators (risk and
reliability of
warnings)
Predictive analytics for proactive terminal process
management
@ duisport inland port terminal
17. Predictive analytics for proactive
process management
BMDA 2018, Vienna 17
monitor
predict
real-time
decision
proactive
management
time
t t +
planned /
acceptable situations
= Violation
= Non-
Violation
e.g., delay in
freight delivery
time
e.g., schedule
faster means of
transport
18. Predictive analytics for proactive
process management
Prediction accuracy key for proactive process management
Prediction accuracy = ability of prediction technique
– to forecast as many true violations as possible,
– while generating as few false alarms as possible
• True violation triggering of required adaptations
– Missed required adaptation = less opportunity for proactively
preventing or mitigating a problem
• False alarm triggering of unnecessary adaptation
– Unnecessary adaptation = additional costs for executing the
adaptations, while not addressing actual problems
BMDA 2018, Vienna 18
19. Predictive analytics for proactive
process management
“Utility” of adaptation decisions depends on…
(1) Accuracy of individual prediction
• Research focused on aggregate accuracy
– E.g., precision, recall, mean average prediction error, …
– But: aggregate accuracy gives no direct information about error of an
individual prediction
Use reliability estimate to quantify probability of violation
BMDA 2018, Vienna 19
Aggregate Accuracy
75%
75%
75%
75%
Prediction #
1
2
3
…
Reliability Estimate
60%
90%
70%
…
20. Predictive analytics for proactive
process management
“Utility” of adaptation decisions depends on…
(2) Severity of violation
– E.g., contractual penalties (such as stipulated in SLAs)
Use estimated penalty to quantify severity (in terms of costs) based
on size of deviation: c()
BMDA 2018, Vienna 20
δ
Linear with cap
clin
c
0 1
δ
Constantc
0
cconst
1
c
Step-wise (s steps)
δ
1/s 2/s
cstep
1
2/s·cstep
1/s·cstep
(s-1)/s
…
0
21. Risk estimate for proactive
process management
Risk =
Probability of occurrence × Severity [ISO 31000:2009]
Reliability estimate × Estimated penalty
BMDA 2018, Vienna 21
22. Risk estimate for proactive
process management
22BMDA 2018, Vienna
Prediction T
Process
Moni-
toring
Data
{
Regression Model 1
Regression Model n
a1
an
{ Deviation Penalty c()
Reliability estimate
Classification Model 1
Classification Model m{{{ Each model of ensemble trained
differently (bagging)
T1
Tm
Ensemble
Prediction:
23. Risk estimate for proactive
process management
BMDA 2018, Vienna 23
monitor
predict
real-time
decision
proactive
management
time
t t +
planned /
acceptable situations
= Violation
= Non-
Violation
R ≤ threshold no adaptation
R > threshold adaptation
+ Risk R
24. Risk estimate for proactive
process management
BMDA 2018, Vienna 24
Costs
Adaptation Cost
Adaptation Cost
+ Penalty
R >
R ≤ No
Adaptation
Adaptation
Risk R
Violation
Non-Violationeffective
not
effective
0
PenaltyViolation
Non-Violation
• Risk threshold
• Adaptation effectiveness
„Utility“ measured in terms of saved costs:
25. Risk estimate for proactive
process management
Initial experimental evaluation
based on air cargo process
5 months of operational data
3 942 process instances
56 082 service invocations
25BMDA 2018, Vienna
Point of
Prediction
26. Constant penalty Nonconstant Penalties
BMDA 2018, Vienna 26
Metzger & Föcker, CAiSE 2017
https://doi.org/10.1007/978-3-319-59536-8_28
Metzger & Bohn, ICSOC 2017
https://doi.org/10.1007/978-3-319-69035-3_25
Cost Savings
Frequency
Cost savings of
14% on average
(in 82.9% of cases)
Additional cost
savings of 23% on
average
Additional Cost Savings
Risk estimate for proactive
process management
27. Agenda
1. TT and the Big Data Value Ecosystem
2. TT Methodology
3. Transport Innovation via Big Data
4. Future Opportunities and Barriers
BMDA 2018, Vienna 27
28. Opportunities
Cross-sector data sharing (e.g., traffic flow passenger flow flights)
Open data: http://europeandataportal.eu/data/en/group/transport
Meta data repositories: e.g., TT Data Portal:
BMDA 2018, Vienna 28
29. Opportunity
Deep Learning
Deep Learning (“AI”)
• Recurrent Neural Networks (RNNs) with LSTM
• Can handle arbitrary length sequences of events
• Initial results for predictive transport process management
– 27% higher accuracy than classical Multi-Layer Perceptron (MLP)
– Robust against how data is encoded (not need to „tweak“)
BMDA 2018, Vienna 29
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0 10 20 30 40 50 60 70 80 90
Diagrammtitel
num s2e noplannednopath mlpCheckpoint [relative prefix]
Accuracy[MCC]
RNN
MLP
30. Barriers
Protection of Personal Data (EU GDPR) – e.g., TT: 1% of TT data
Protection of Commercial Data / IPR – e.g., 68% of TT data sources
BMDA 2018, Vienna 30
https://www.eugdpr.org/
http://www.industrialdataspace.org/en
31. Barriers
Lack of skills
• Demand for “data professionals”
exceeds supply on the labour market
• (Some) ongoing activities
– EU “Digital Skills and Jobs” Coalition
– BDV PPP
Education Hub:
(Sources: [OECD, 2015; IDC 2015])
Year Gap (total
EU)
Gap (% EU)
2014 500,000 8%
2020 (baseline) 530,000 6%
2020 (challenged) 150,000 2%
2020 (high-growth) 3,500,000 30%
BMDA 2018, Vienna 31
32. Thank You!
BMDA 2018, Vienna 32
This project has received funding from the European Union’s Horizon 2020
research and innovation programme under grant agreement no. 731932
33. Mobility and Logistics
One of most-used industries in the world and in EU…
• 15% of GDP (source: KLU)
• Employment of 11.2 million persons in EU-28 (source: DG MOVE)
• 3,768 billion tonne-kilometres and 6,391 billion person-kilometres in EU-28
• Key contributor to emissions: 4,824 megatonnes CO2 (source: DG MOVE)
…and growing
• Business and tourism travel expected to grow significantly over next decades
• Freight transport slated to increase by 40 % in 2030 and by 80% in 2050
(source: ALICE ETP)
Need for paradigm shift!
• 10% efficiency improvement = EU cost savings of 100 B€ (source: ALICE ETP)
• Big Data expected to lead to 500 billion USD in value worldwide in the form
of time and fuel savings, and savings of 380 megatons CO2 in transport and
logistics (source: OECD)
BMDA 2018, Vienna 33
Notas do Editor
1,16 MEUR für paluno
1,16 MEUR für paluno
Rechts: hist(d$nonconstant, breaks = 20, density = 20, ylim=c(0,40), xlim=c(.1,.35))
OECD: 2015_final_OECD-Datadriven_Innovation
IDC: EDM_D6_Interim Report Release October 16 2015_Final