flux vision - digitaltransport4africa · 8 orange business services data anonymisation 1 2 3 the...
TRANSCRIPT
1 Orange Business Services
Flux VisionAFD
November 9th 2017
Vers une plateforme de données numérique
collaborative et ouvertes pour améliorer le transport
public urbain en Afrique
2 Orange Business Services
SUMMARY
Introduction
1.Floating Mobile Data for Transportation: 3 main challenges
2.Orange Flux Vision Solution
3.Project with AFD – CETUD (Senegal) and perspectives in Africa
4 Orange Business Services
Flux VisionReal time statistics on mobility patterns
Flux Vision converts in real time mobile network data into statistical indicators
Commercially mature product launched in several countries
Massive data inflow handling
• Capability to incorporate any kind of location data
Privacy compliant by design
• Irreversible anonymization through approved algorithms
A solution entirely designed by Orange
• 8 years and more of R&D
• Continuous technical roadmap
An extensive background in Radio Modelling and Coverage
• Accurate mobile signaling data localization
• Worldwide simulation expertise (>250 MNO’s)
5 Orange Business Services
Network Simulation: Spatial Uncertainity
Geo referencing Multiple technologies
GSM
Edge
LTE
6 Orange Business Services
Mobile Signaling: Temporal Uncertainity
Mobile 1
timeline
Mobile 2 Mobile 3
T0 T0+TS T0+2TS
Export of all detections
performed in the interval [T0;
T0+TS] to the data flow.
Export of all detections performed
in the interval [T0+TS; T0+2TS]
to the data flow.
Data Aggregation
T0+3TS
Mobile 1Mobile 2 Mobile 3Mobile 2
Detection performed Detection performed Detection performed
Mobile 1 detected
Mobile 2 detectedMobile 1 detected
Mobile 3 detected
Export of all detections performed
in the interval [T0+2TS; T0+3TS]
to the data flow.
Mobile 2 detected
Mobile 3 detected
No detection performed
7 Orange Business Services
Observable target behavior
Invisible Target behavior
1 2Rectify spatial-
temporal biases that are specific to the
mobile network
Extrapolate the number of mobiles to
the "standard" population
3Rectify the volume of "standard" people according to the observed
context
Counting considers
▪ Mobile clients from Orange▪ Foreign Sim cards
Roaming in Orange’s Network
To extrapolate the total population present on the area, rectification algorithms are applied to the collected data:
1. To compensate mobiles present but not captured2. To represent the "standard" population present in the area3. To avoid over or under-representations depending on the context
Data fine tuning and extrapolation
8 Orange Business Services
Data Anonymisation
1
2
3
The application of the text will be mandatory on 25 May 2018 in all the countries of
the European Union.
Processing of personal data such as pseudonymisation and profiling requires prior
consent from the data subject or some other legitimate basis.
Financial sanctions are foreseen for any breach of the regulation, the highest up to
4% of the annual global turnover.
The Regulation does not apply in the case of data that have been anonymized. 4
GDPR (General Data Protection Regulation)
10 Orange Business Services
Some of our references
A mature commercial solution with more than 150 business customers
TransportationEvents
Retail and geomarketing Tourism
11 Orange Business Services
• Measure seasonal trends
• Sizing of logistical resources
• Study the impact of a particular event
Dynamics of travel demand
Exploitation Geomarketing and transports
Deep knowledge of transports demand
• Evaluate the pertinence of the existing transportation offer
• Plan the evolution of the infrastructure
• Feed mobility models
• Evaluate the attractiveness of my infrastructures
• Determine catchment areas
• Study Market shares and competing modes or networks
• Collect, analyze and verify real-time traffic data.
• Gather better information to better react
FMD for Transportation4 main axes around Transportation
12 Orange Business Services
Big Data, large and representative samples
Revealed preferences, in opposition to declaration :limited bias, but also limited information
Spatial temporal precision
Full Measure – all the time, everywhere
Modal split and trip purpose limitations
No additional investment, lower costs
Strengths Weaknesses
Strenghts and Weaknesses
13 Orange Business Services
What we deliver
Data visualizationTabular Files
Csv files
Somme de Volume Étiquettes de colonnes
Étiquettes de lignes 1 2 3 4 5 6 7 8 9 10 11
13/03/2017
1 1961 60 101 0 69 153 0 0 41 0
2 40 1794 50 0 0 91 0 0 0 0
3 132 50 3893 248 0 132 0 83 52 61 82
4 0 0 213 3971 0 0 40 62 104 41 133
5 0 0 0 0 1761 102 51 135 0 0 0
6 132 120 101 0 99 2758 0 0 0 0
7 0 50 0 0 99 0 2194 156 0 61
8 0 0 0 69 112 81 2953 0 133
9 0 71 0 0 0 0 0 1902 82 71
10 0 0 0 0 0 0 0 41 3093 174
11 0 0 0 0 0 218 62 164 4419
12 0 0 39 0 0 0 41 0 41 174
13 0 0 0 0 0 0 102 0
14 40 0 0 0 169 51 153 145 0 0 41
15 0 0 0 0 0 0 0 62 0 0
16 0 0 0 69 0 0 0 0 52 339 154
17 0 0 0 0 0
18 0 0 0 0 0 41
19 0 60 0 0 0 0 0 0 113 133
20 0 0
21 0 0 0 0 0
22 0 0 0 0 40 0 0 0 71
23 0 40 50 142 0 40 40 41 0 133 184
24 0 0 71 0 0 0 0 72 0 71 123
25 0 60 0 0 49 0 51 62 41 71 113
26 0 0
27 0 0 0 0 0 0
28 0 0
29 0 0
30 0
Total 13/03/2017 2305 2174 4610 4469 2315 3439 2650 4030 2254 4352 6107
14/03/2017
1 2002 39 131 0 80 144 0 0 0 0 0
2 0 1843 91 0 0 124 0 0 0 0 0
3 163 49 3851 233 0 124 62 72 62 103 92
4 71 49 326 3784 90 62 0 93 134 82 206
5 0 0 0 1727 62 0 114 0 61
6 102 149 71 50 110 2484 0 93 0 0 0
7 0 0 0 0 90 62 2245 124 0 82
8 40 0 0 50 80 82 103 2793 41 0 134
9 0 81 50 0 0 0 51 1708 82 92
10 0 0 0 0 51 0 41 124 3020 288
11 0 0 0 0 0 41 0 145 0 144 4379
12 0 0 0 60 0 41 51 103 0 82 355
13 0 0 0 0 0 0 0 154 51
14 40 0 0 0 70 51 134 134 0 0 41
15 0 0 0 0 0 51 41 0 51 0
16 0 0 40 70 0 0 0 114 433 175
17 0 0
18 0 0 0 0 0 0 0 0 0
19 0 0 0 0 0 0 0 0 175 159
20 0 0 0
21 0 0 0 0 0
22 0 0 0 0 0 0 0 0 0 0 72
23 0 0 50 136 0 82 82 41 0 92 185
24 0 39 81 0 0 0 41 0 41 82 195
25 0 49 0 70 60 0 0 83 62 82 61
26 0
27 0 0 0 0 0 0
OD Easily handled with Excel Pivot Tables
16 Orange Business Services
Enjeux / Plus value apportée à Dakar
1
2
3
Actualiser les données de mobilité à Dakar et ainsi faciliter la planification des transports
sur le territoire :
• Compléter la connaissance des déplacements des franges périurbaines ;
• Connaître les évolutions de la mobilité lors d’événements spécifiques (ramadan,
inondation) ;
• Estimer les effets de la mise en place de la gare routière des Baux Maraîchers.
Établir le domaine de pertinence des données mobiles et les méthodologies qui
permettent de les intégrer dans des contextes similaires au Sénégal et ailleurs.
Mettre en place une méthodologie pour la tenue à jour de la connaissance de la mobilité
à Dakar entre deux enquêtes ménages déplacements (EMD) (intervalle 10-15 ans)
Participants
17 Orange Business Services
Origin Destination Matrix – Strates de Dakar
Area Visitor’s Origin
Angle Mousse 8358
Bambilor Dani(Nord) 1556
Bargny Extension (Est) 1666
Camberene Centre 7836
Camp le Clerc-Liberte VI Extension 26587
Champ de courses Arafat Gouye Mouride 6249
Cites Residentielles Littorales 12702
Dalifort 54062
Dieupeul 6364
Doro Aw 8744
DTK Ouest 18442
Extension Nord Ouest 4032
Fann Residence 7126
Fith Mith 14747
Grand Dakar Sud 3368
Gueule Tapee 7215
HLM 5-6 (Centre) 6227
Hors zone 4179
ICOTAF SOTIBA 55001
KM Ainoumady 16515
Catchment Area of the new Bus Station: Baux Maraîchers
Area An
gle
Mo
uss
e
Bam
bilo
rD
ani(
No
rd)
Bar
gny
Exte
nsi
on
(Est
)
Cam
be
ren
eC
en
tre
Cam
p L
e C
lerc
Ch
amp
de
C
ou
rse
s
Cit
es
Re
sid
en
tie
lles
Dal
ifo
rt
Die
up
eu
l
Do
ro A
w
DTK
Ou
est
Exte
nsi
on
No
rd
Ou
est
08/09/2015
Angle Mousse 42 38 213 639 44 123 530 171 1071 1257 34
Bambilor Dani(Nord) 0 110 0 151 799 71 84 114 23 55 692
Bargny Extension (Est) 0 149 0 101 963 60 93 34 0 29 127
Camberene Centre 162 86 57 1475 84 117 397 265 484 219 36
Camp le Clerc-Liberte VI Extension 232 187 134 440 237 368 1022 5040 318 412 187
Champ de courses Arafat Gouye Mouride 0 616 664 0 215 158 230 88 42 62 3639
Cites Residentielles Littorales 55 92 103 55 454 232 1119 136 97 285 103
Dalifort 93 68 77 69 773 123 462 311 143 316 88
Dieupeul 34 44 31 81 3230 79 75 206 69 95 49
Doro Aw 705 58 33 524 788 82 191 663 206 2296 51
DTK Ouest 1399 138 93 296 942 228 458 1698 322 2606 112
Extension Nord Ouest 0 828 186 33 289 6112 108 163 79 0 68
0 10 000 20 000 30 000 40 000 50 000 60 000
Bambilor Dani(Nord)
Grand Dakar Sud
Hors zone
Champ de courses Arafat Gouye Mouride
Fann Residence
Camberene Centre
Doro Aw
Fith Mith
DTK Ouest
Dalifort
18 Orange Business ServicesOrange Footprint as Mobile Operator + Flux Vision as a Service to other Operators
19 Orange Business Services
Thank you!
For more information…
Marcelo Pimont Strambi
Flux Vision International Project Manager
22 Orange Business Services
Historical CEM (Customer Experience Management) approach
In a complex environment…
23 Orange Business Services
Statistical IndicatorsFlux Vision PlatformMobile Network
Flux Vision : how does it work?
Blurring, Hashing, Aggregation and Anonymization of
data on the fly
Technical fine-tuning and
extrapolation to the total population
Date Jour de la semaine Pays N Departement Continent Departement Region Visiteur Uniques sur Zone Analyse Categorie Visiteur Volume
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 34
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Mois Centre Commercial 1 Resident 37
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées 3Mois Centre Commercial 1 Resident 37
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 1 Resident 22
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 23
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 21
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 25
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 25
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Mois Centre Commercial 1 Resident 30
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées 3Mois Centre Commercial 1 Resident 30
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 1 Resident 26
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 31
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Mois Centre Commercial 2 Resident 25
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées 3Mois Centre Commercial 2 Resident 25
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 32
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 24
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Mois Centre Commercial 1 Resident 21
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées 3Mois Centre Commercial 1 Resident 21
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 24
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 21
11/11/2016 Vendredi France 31 Europe Haute-Garonne Midi-Pyrénées Jour Centre Commercial 2 Resident 21
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu Jour Centre Commercial 1 TouristeEtranger 54
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu Jour Centre Commercial 2 TouristeEtranger 82
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu Mois Centre Commercial 1 TouristeEtranger 47
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu Mois Centre Commercial 2 TouristeEtranger 47
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu 3Mois Centre Commercial 1 TouristeEtranger 47
11/11/2016 Vendredi Allemagne Inconnu Europe Inconnu Inconnu 3Mois Centre Commercial 2 TouristeEtranger 47
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu Jour Centre Commercial 1 TouristeEtranger 41
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu Jour Centre Commercial 2 TouristeEtranger 42
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu Mois Centre Commercial 1 TouristeEtranger 25
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu Mois Centre Commercial 2 TouristeEtranger 36
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu 3Mois Centre Commercial 1 TouristeEtranger 25
11/11/2016 Vendredi Belgique+Luxembourg Inconnu Europe Inconnu Inconnu 3Mois Centre Commercial 2 TouristeEtranger 36
Generic Tabular
Data
Interactive BI Tools
Origine / Destinations : examples
Origins distribution to area 14 Destinations distribution of people having
spent the night in area 10
Cartographic Analysis
Real-time data stream
Probes or CDR
CRM Database
Additional segmentations from the
MNO’s line of business data
Behavior Segmentation
Definitions
Network Simulation
Results
The system requires to know precisely the
analytics to be computed before data processing
24 Orange Business Services
Other Data Sources that we might use in our studies
Wi-Fi logs▪ Interface with the WIFI provider ▪ Geolocation of the visitors ▪ Creation of precise study areas
Mobile Apps GPS DataMultiple Application geolocated data
Qualitative research dataDeclarative Information via Apps
On-Site MeasuringAd-hoc testing of network communication
Open DataEx: Weather Data - Attendance impact and correlation analysis
Private DataQuantaFlow, Parking, Cameras…etc
Smart Cities and IOTConnected Devices data
25 Orange Business Services
Deep knowledge of transport demandInfrastructure and transport network planning
Some referencesKey elements of the technical solution
New functionalities
Link norm-atisLink BI
Brussels
Extrapolation models enrichmentprofessional mobiles / internationalroaming / M2M sim cards
Feedback for contextualization according to transport infrastructures
Main stakes• Estimation of the end to end demand: better
size the existing networks and optimize their
exploitation.
• Large samples. Reduced Costs.
• Flexibility on the analysis zoning
• Feed mobility models
• Accessibility and Environmental impact studies
Origin-Destination Matrix
• Territory divided in Origin and Destination zones• Multiple partitions, independent from administrative divisions.• Several levels of analysis (micro/macro)
• Volume of journeys from an origin to a destination with an immobility time
• Main overnight stay and activity area of the visitors of an area
• Segmentation of travel volumes by• Criteria that are measured before and during the movement
(modes of transport, travel purpose)• Criteria that are independent of the observed displacements
(socio-demographic)
Objectives
✓ Complementary data for new infrastructure development
✓ Modal share studies
✓ Mobility indicators standardization
Clients
Transport operators, Public transportation boards, Public authorities, Infrastructure Constructors and Concessions
26 Orange Business Services
Dynamics of travel demandWhich seasonal impacts, following external events or as a result of changes in supply (new lines, construction work, etc.)
Main Stakes• Dimension logistic resources
• Feed prediction models
• Evaluate the impact on OD or modal shift dueto changes in transport supply
• Review and organize line schedules
• Negotiate with public and private actors usinga more precise basis
More frequent and richer mobile network locator eventsIncreased spatio-temporal accuracy and wider sample
Automated post-processingResults delivery in near real-time(France)
Evolution of indicators over time
• Day-to-day volumes of OD trips
• For any study period, or even permanently
• Presence and trips hour by hour
• According to time of arrival at destination
• Evolution of visitor / traveler profiles:
• Trip purpose (tourists, workers), origin, duration of stay
Some referencesKey elements of the technical solution
New functionalities
Objectives
✓ Station access planning
✓ Impact due to construction work in railway
✓ Impact of major events
Clients
Public transport authorities, Local authorities, Event organizers
27 Orange Business Services
ExploitationCollect, analyze and verify traffic data. Inform for action
Real-time counting
• Focus on transportation infrastructures and lines
• 1/30 minutes attendance of stations, lines, roads, railway sections and various places
• Inbound / Outbound Volumes
• Real-time monitoring of mobile groups moving in a coherent way on a rail network:
• Virtual beacons: passing, identification and passage counting
• Integration into your environment• KPI push by sFTP, csv or xml files
Main Stakes• Provide real-time reliable indicators.
• Warnings - Traveler Information
• Monitor service operations
• Adapt to the gradual rise of competition.
More frequent and richer mobile network locator eventsIncreased spatio-temporal accuracy and wider sample
Automated post-processingResults delivery in near real-time(France)
Enrichment with other data sources (GPS, Operations, etc.)
Some referencesKey elements of the technical solution
New functionalities
Objectives
✓ Inbound / Outbound Underground Stations
✓ 15 minutes attendance of a Train Station
✓ Tracking of regional trains
Clients
Railway operators, Public transportation boards
28 Orange Business Services
Geomarketing et transportsEvaluer l’attractivité de mes infrastructures
Catchment Areas and Competitive Analysis
• Territory cut out in customized isochrones• areas of interest / areas of origin
• Understanding the profile and behavior of travelers• Origin• Duration of Presence• Purpose• Socio-demo
• Evaluation of the impact of multi-media advertising campaigns
New Standard Offers
BI and Data Visualisation
Indoor Analysis (Wi-Fi, Dedicated Antennas)
Main stakes• Knowing the catchment areas of a railway
station, airport, etc.
• Study your market share, penetration rate and
competing modes or networks
• Segmentation of observed populations and
valuation of commercial spaces
• Integrate new services (car-sharing, car-pooling,
bicycling, etc.) with a global approach at stations
and transport nodes
• Valuation of the rental or advertising spaces of
the nodes / axis of a transport infrastructure
Distribution of passengers arriving at the airport into the multiple train stations
Some referencesKey elements of the technical solution
New functionalities
Objectives
✓ Place of residence of the travelers of an Airport
✓ Comparison of catchment areas between modes
✓ Impact of targeted publicity campaign
Clients
Airports, Road / Rail operators, Infrastructure managers