39
Social Network Analysis using Mobile Phone Data António Pedro Oliveira University of Coimbra, Portugal

Social network analysis using Mobile phone data

Embed Size (px)

Citation preview

Page 1: Social network analysis using Mobile phone data

Social Network Analysis using Mobile Phone Data

António Pedro OliveiraUniversity of Coimbra, Portugal

Page 2: Social network analysis using Mobile phone data

Introduction

Motivation Social networks

Social media (facebook, twitter, linkedin) Obtain, web crawling of data Find relations among people Understand relations’ type

2

Introduction Research Areas Data Tools Conclusion

Page 3: Social network analysis using Mobile phone data

Research Areas Social Network Analysis

Software tools Gephi

Nodes Edges Adjacency matrix Node degree

3

Introduction Research Areas Data Tools Conclusion

Page 4: Social network analysis using Mobile phone data

Research Areas Social Network Analysis

Centrality of a node in degree (number of incoming edges), out degree (number of outcoming edges), betweenness (number of pairs of nodes

that go through you) closeness (average distance of one node to

anothers)

4

Introduction Research Areas Data Tools Conclusion

Page 5: Social network analysis using Mobile phone data

Research Areas Social Network Analysis

Community mutuality of ties (everybody in the group knows

everybody else) frequency of ties among members (everybody in

the group has links to at least k others in the group)

cliques (every member of the group has links to every other member)

clustering

5

Introduction Research Areas Data Tools Conclusion

Page 6: Social network analysis using Mobile phone data

Research Areas

Social Network Analysis Finding motifs (small structures) in

network models Average shortest path

6

Introduction Research Areas Data Tools Conclusion

Page 7: Social network analysis using Mobile phone data

Data

7

Introduction Research Areas Data Tools Conclusion

Two main files cell_dim and call_fct cell_dim file presents longitude, latitude

and regions for each cell tower (1) cell_id (5) longitude (6) latitude (7) region = region ID (There are 3 regions)

Page 8: Social network analysis using Mobile phone data

Data

8

Introduction Research Areas Data Tools Conclusion

call_fct file presents caller ID and callee ID and respective cell towers, as well as the date and duration of the call

(1) originating_id (caller ID) (2) originating_cell_id (caller's cell tower) (3) terminating_id (callee ID) (4) terminating_cell_id (callee's cell tower) (5) date_id  (6) duration_amt (duration of the call)

Page 9: Social network analysis using Mobile phone data

Data

9

Introduction Research Areas Data Tools Conclusion

lisbon_subject_id file has the subjects’ ids of Lisbon which helps us to identify phone calls of Lisbon

fct_call_oneDay is a file obtained with the phone calls for only one day

Page 10: Social network analysis using Mobile phone data

Using Mobile Phone Data

10

Introduction Research Areas Data Tools Conclusion

Which opportunities we find on cellphone data of OPTIMUS

Grouping people Location of people People mobility (means of transport, POIs) Duration of phone calls Date

Page 11: Social network analysis using Mobile phone data

Tools

11

Introduction Research Areas Data Tools Conclusion

Gephi Clustering algorithm (MCL)

Divide the social network in clusters of points Some of the various statistics parameters

Average degree Network diameter Graph density Average clustering coefficient Average path length

Page 12: Social network analysis using Mobile phone data

Tools

12

Introduction Research Areas Data Tools Conclusion

Gephi (fct_max_300(id))

Page 13: Social network analysis using Mobile phone data

Using Mobile Phone Data

13

Introduction Research Areas Data Tools Conclusion

Select data of one day, where the caller is in a specific location (for instance, in Lisbon) Obtain a social network For the ones with the longer duration’s

calls represent the location in QGIS

Page 14: Social network analysis using Mobile phone data

Tools

14

Introduction Research Areas Data Tools Conclusion

Gephi (fct_lisbon_oneDay)

Page 15: Social network analysis using Mobile phone data

Tools

15

Introduction Research Areas Data Tools Conclusion

Postgres To select cell towers from LisbonSELECT id, latitude, longitude

FROM cell_dim

WHERE latitude < -9.10 AND latitude > -9.42

AND longitude > 38.69 AND longitude < 38.96

LIMIT 10000

Page 16: Social network analysis using Mobile phone data

Tools

16

Introduction Research Areas Data Tools Conclusion

Postgres To select phone calls of one day with cell

towers belonging to Lisbonselect *

from fct_call

where date_id >= 35100000 AND date_id < 35200000

AND (originating_cell_id = 10011 OR …)

Page 17: Social network analysis using Mobile phone data

Tools

17

Introduction Research Areas Data Tools Conclusion

QGIS (longer phone calls in Lisbon)

Page 18: Social network analysis using Mobile phone data

Tools

18

Introduction Research Areas Data Tools Conclusion

Use previous data and for each hour in a day identify subjects with the higher duration of calls Present the variation for each hour in

QGIS

Page 19: Social network analysis using Mobile phone data

Tools

19

Introduction Research Areas Data Tools Conclusion

QGIS (hour 0)

Page 20: Social network analysis using Mobile phone data

Tools

20

Introduction Research Areas Data Tools Conclusion

QGIS (hour 1)

Page 21: Social network analysis using Mobile phone data

Tools

21

Introduction Research Areas Data Tools Conclusion

QGIS (hour 2)

Page 22: Social network analysis using Mobile phone data

Tools

22

Introduction Research Areas Data Tools Conclusion

QGIS (hour 3)

Page 23: Social network analysis using Mobile phone data

Tools

23

Introduction Research Areas Data Tools Conclusion

QGIS (hour 4)

Page 24: Social network analysis using Mobile phone data

Tools

24

Introduction Research Areas Data Tools Conclusion

QGIS (hour 5)

Page 25: Social network analysis using Mobile phone data

Tools

25

Introduction Research Areas Data Tools Conclusion

QGIS (hour 6)

Page 26: Social network analysis using Mobile phone data

Tools

26

Introduction Research Areas Data Tools Conclusion

QGIS (hour 7)

Page 27: Social network analysis using Mobile phone data

Tools

27

Introduction Research Areas Data Tools Conclusion

QGIS (hour 8)

Page 28: Social network analysis using Mobile phone data

Tools

28

Introduction Research Areas Data Tools Conclusion

QGIS (hour 9)

Page 29: Social network analysis using Mobile phone data

Tools

29

Introduction Research Areas Data Tools Conclusion

QGIS (hour 10)

Page 30: Social network analysis using Mobile phone data

Tools

30

Introduction Research Areas Data Tools Conclusion

QGIS (hour 11)

Page 31: Social network analysis using Mobile phone data

Tools

31

Introduction Research Areas Data Tools Conclusion

QGIS (hour 12)

Page 32: Social network analysis using Mobile phone data

Tools

32

Introduction Research Areas Data Tools Conclusion

QGIS (hour 13)

Page 33: Social network analysis using Mobile phone data

Tools

33

Introduction Research Areas Data Tools Conclusion

QGIS (hour 14)

Page 34: Social network analysis using Mobile phone data

Tools

34

Introduction Research Areas Data Tools Conclusion

QGIS (hour 15)

Page 35: Social network analysis using Mobile phone data

Tools

35

Introduction Research Areas Data Tools Conclusion

QGIS (hour 16)

Page 36: Social network analysis using Mobile phone data

Tools

36

Introduction Research Areas Data Tools Conclusion

QGIS (hour 17)

Page 37: Social network analysis using Mobile phone data

Conclusions Challenges

Obtain semantic information from mobile phone data

Understand the meaning of the (groups in) social networks (friendship?)

Understand geographic information Relate social networks with geographic information The mobility of the users (POIs?) Obtain profile for the users

37

Introduction Research Areas Data Tools Conclusion

Page 38: Social network analysis using Mobile phone data

Conclusions

We are using mobile phone data of OPTIMUS to find social networks

We obtained social networks with small groups (of interactions)

We showed the distribution of mobile phone data on a GIS

38

Introduction Research Areas Data Tools Conclusion

Page 39: Social network analysis using Mobile phone data

The End

39