Mobile phone data for migration analysis: new possibilities, risks and difficulties. Zbigniew Smoreda SENSE (Orange Labs, Paris)

Size: px

Start display at page:

Download "Mobile phone data for migration analysis: new possibilities, risks and difficulties. Zbigniew Smoreda SENSE (Orange Labs, Paris)"

Melanie Arline Norton
5 years ago
Views:

Paris) KCMD-GMDAC workshop o Big Data ad alterative data

1 Mobile phoe data for migratio aalysis: ew possibilities, risks ad difficulties Zbigiew Smoreda SENSE (Orage Labs, Paris) KCMD-GMDAC workshop o Big Data ad alterative data sources o migratio: from case studies to policy support, Ispra, 30 Dec. 2017

CDRs telcos billig records Billig data collected automatically for all MPO s customers timestamp party A party B type duratio cell A cell B 16/10/30

2 CDRs telcos billig records Billig data collected automatically for all MPO s customers timestamp party A party B type duratio cell A cell B 16/10/30 10:01: SMS /10/30 10:01: Voice /10/30 10:01: Voice

3 CDRs telcos billig records Billig data collected automatically for all MPO s customers timestamp party A party B type duratio cell A cell B 16/10/30 10:01:33 ABCXRTTAAT XYRAATRRAA SMS 0 Lat, Lo Lat, Lo 16/10/30 10:01:34 ZXXRTT554R CC445EERSDA Voice 54 Lat, Lo Lat, Lo 16/10/30 10:01:35 776EREREER 99TRDDAAA7 Voice 132 Lat, Lo Lat, Lo 3

SMS 0 Lat, Lo Lat, Lo 16/10/30 10:01:34 ZXXRTT554R CC445EERSDA Voice 54 Lat, Lo Lat, Lo

4 CDRs telcos billig records Billig data collected automatically for all MPO s customers timestamp party A party B type duratio cell A cell B 16/10/30 10:01:33 ABCXRTTAAT XYRAATRRAA SMS 0 Lat, Lo Lat, Lo 16/10/30 10:01:34 ZXXRTT554R CC445EERSDA Voice 54 Lat, Lo Lat, Lo 16/10/30 10:01:35 776EREREER 99TRDDAAA7 Voice 132 Lat, Lo Lat, Lo Oela et al Gozàlez et al. 2008

5 Quickly growig research 5

6 NetMob commuity buildig 6

7 NetMob commuity buildig 7

8 How to accelerate ad ope access to CDRs? A Ope Iovatio Project with Orage Cote d Ivoire, Orage Marketig Visio ad Corporate Social Resposibility Data for Developmet (D4D) Cote d Ivoire

D4D: data for developmet Cote d Ivoire 5 moths of deidetified CDRs OCI ateae positios Orage Data Ceter data processig at Orage Labs Proposed datasets: (1) atea-to-atea traffic o a hourly basis, (2)

9 D4D: data for developmet Cote d Ivoire 5 moths of deidetified CDRs OCI ateae positios Orage Data Ceter data processig at Orage Labs Proposed datasets: (1) atea-to-atea traffic o a hourly basis, (2) idividual trajectories for 50,000 radomly sampled users for two week time widows with atea locatio iformatio, (3) idividual trajectories for 500,000 radomly sampled users over the etire observatio period with sub-prefecture locatio iformatio, ad (4) a sample of commuicatio graphs for 5,000 customers 9

10 D4D: data for developmet Cote d Ivoire Scietific challege - halfway betwee a log hackatho ad a scietific coferece: Lauched i Jue 2012 for 8 moths oly Oly research istitutios admitted after sigig the terms & coditios Evaluatio committee chaired by Vicet Blodel (UCL) with members from Bouake Uiversity, Global Pulse (UN), GSMA, Orage Labs, WEF ad MIT Selected projects presetatio durig the NetMob coferece, May 2013 i Bosto (USA) Four prizes attributed (first prize, scietific, developmet, ad data visualizatio prizes) 10

11 D4D: data for developmet Côte d'ivoire CDR samples Ope to the researchers 263 projects submitted Over 80 reports received 11

D4D: data for developmet Cote d Ivoire http://perso.

12 D4D: data for developmet Cote d Ivoire 12

13 D4D: data for developmet Cote d Ivoire + 50 published works! 13

14 D4D: data for developmet Cote d Ivoire May various topics addressed by the projects: Health improvemet (disease spread mappig or prevetio) Populatio statistics (urbaizatio, populatio, tourism ad evets aalysis) Commuities uderstadig (diaspora cartography ad eeds, rural ad urba customers) Ecoomic Idicators (local ecoomic developmet, micro fiace isight) City ad trasport plaig (trasport optimizatio, road costructio, smart city plaig) Emergecy, Alertig & Prevetig (early warig system, help distributio localizatio) Geo-marketig (strategic poits of sales) Importat echo amog the Uited Natios, NGOs, developmet aid istitutios New data challeges iitiatives (Telecom Italia, Telefoica) But o project could be implemeted i Ivory Cost 14

15 Secod D4D Seegal: do it differetly

16 D4D Seegal Improvemets to brig to D4D after the Cote d Ivoire experiece 1. Focus o 5 themes Health, Trasport, Agriculture, Eergy, Natioal statistics 2. Ivolve local ecosystem Questios owers: Miistries ad istitutios Cotributors: Uiversities, etrepreeurs 3. Reiforce goverace Regulatio, Ethics, 4. Foster Data sharig Fid other data sources to be crossed with D4D data sets 5. Esure a implemetatio of results Prepare the «After D4D» 16

D4D Seegal 12 moths of deidetified CDRs Soatel ateae positios Orage Data Ceter data processig at Orage Labs Proposed datasets: (1) atea-to-atea traffic o a hourly basis, (2) idividual trajectories

17 D4D Seegal 12 moths of deidetified CDRs Soatel ateae positios Orage Data Ceter data processig at Orage Labs Proposed datasets: (1) atea-to-atea traffic o a hourly basis, (2) idividual trajectories for 300,000 radomly sampled users for two week time widows with atea locatio iformatio, (3) idividual trajectories for 150,000 radomly sampled users over the etire observatio period with arrodissemet locatio iformatio, ad (4) a sample of commuicatio graphs for 5,000 customers Badicoot idividual idicators with (2) & (3) 17

18 D4D Seegal April 2014: challege lauched 18

19 D4D Seegal results aoucemet 19

20 the wiers 20

21 Seegal iteral migratios Marti-Gutierrez et al. (2016) Agricultural activity shapes the commuicatio ad migratio patters i Seegal, Chaos

22 D4D Seegal Results are published o the Web: D4D scietific cotributios ios_scietific_papers.pdf (298 pages, 83MB) D4D posters ios_posters.pdf (56 pages, 53 MB) The datasets for D4D are described here: 22

23 Thaks to B&M Gates Foudatio grats three projects were lauched to explore further the use of CDRs for: Health: Disease modelig Natioal Statistics: Proxy for idicators Agriculture: Food security

24 D4D Seegal: Grats for implemetatio Health Natioal Statistics Agriculture Ucoverig the impact of huma mobility o schistosomiasis... R Casagradi et al., Politecico di Milao, Italy Costructio of sociodemographic idicators with digital breadcrumbs F Brucksche & T Zbiraski, Freïe Uiversität Berli, Germay Geesis of millet prices i Seegal: the role of productio, markets ad their failures D Jacques, UC Louvai, Belgium Quatifyig effect of movemet due to holidays o malaria prevalece A multidimesioal aalysis of poverty ad its determiats i Seegal Mobility profiles ad caledars for food security ad livelihoods aalysis S Milusheva, Brow Uiversity, USA N Pokhriyal, State Uiversity of New York at Buffalo, USA PJ Zufiria et al., Uiversidad Politécica de Madrid, Spai (See: Aex) grated access to the complete database 24

25 D4D Seegal: 6 projects implemetatio Special evet i Dakar ( Radisso Soatel evet, 6 th Jue 2015 ) ad teams field visit to Seegal Collaboratios with local actors: ANSD (Natioal Statistics) Miistry of Health World Food Program Dakar DAPSA Seegal (Agriculture Statistics) PNLP (Natioal Malaria Cotrol Program) Help of iteratioal developmet bodies (OECD, WFP, WHO, UN ) Ad fially a «twi project» o the Natioal Statistics Seegal - Cote d Ivoire is about to start 25

26 After D4D: the world of developmet eeds Big Data DATA-POP The Law, Politics ad ALLIANCE Ethics of Cell Phoe White Paper Data Aalytics April

27 Privacy warig: impossible CDRs aoymizatio The otio of aoymity is oe of the corerstoe of data protectio But a idividual spatiotemporal trace becomes very quickly uique, eve i the largest datasets CDRs from a Europea coutry (1.5 M people, 15 moths) Poits: atea tower / time step: oe hour 27

28 The privacy-utility trade-off Privacy Utility 28

29 A paradigm-shift i data protectio Provide access to richer but hidde data Privacy secure platform with pseudoymized data brig the code to the data usig APIs create a bak of ope algorithms set-up a goverace board (local ad global) Utility 29

30 OPAL Project: query private data i a safe way 30

31 OPAL project The Official Lauch of the OPAL Project took place o 8 December 2016 durig the Ope Govermet Partership Global Summit i Paris 31

Migratio studies specific limitatios of CDRs CDR based studies are usually o a sigle coutry whe they chage coutry, migrats also chage operator (oe eeds to

32 Migratio studies specific limitatios of CDRs CDR based studies are usually o a sigle coutry whe they chage coutry, migrats also chage operator (oe eeds to follow the device IMEI to observe cotiuously their movemets) i Roamig situatio the observatio becomes more radom (device ca switch betwee operators whe abroad) 32

Calibrating Car-following Model with Trajectory Data by Cell Phone

Calibrating Car-following Model with Trajectory Data by Cell Phone Calibratig Car-followig Model with Trajectory Data by Cell Phoe Jiag Zeyu 1 jiagzy15@mails.tsighua.edu.c Wag Yog 3 13006373891@qq.com Wag Gagqiao 1 wag-gq15@mails.tsighua.edu.c Zhag Xiagpeg 3 364386058@qq.com