Your SlideShare is downloading. ×
Understanding social connections
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Understanding social connections

75

Published on

Mobile network big data is used to explore the social interaction patterns in Sri Lanka

Mobile network big data is used to explore the social interaction patterns in Sri Lanka

Published in: Data & Analytics
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
75
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
2
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. This  work  was  carried  out  with  the  aid  of  a  grant  from  the  Interna5onal  Development  Research  Centre,  O>awa,  Canada.     Understanding  social  connec/ons   Kaushalya  Madhawa,  LIRNEasia   Responsible  use  of  mobile  meta-­‐data  to  support  public  purposes   Jetwing  Lagoon,  Negombo   08  August  2014    
  • 2. Understanding social connections •  Can  we  iden5fy  communi5es  of  people  using   social  connec5ons?   •  Do  communica5on  pa>erns  reveal  the  socio-­‐ economic  status  of  a  popula5on?   •  Are  people’s  communica5on  and  mobility   pa>erns  co-­‐related?   1
  • 3. Understanding social connections •  Can  we  iden5fy  communi5es  of  people  using   social  connec5ons?   •  Do  communica5on  pa>erns  reveal  the  socio-­‐ economic  status  of  a  popula5on?   •  Are  people’s  communica5on  and  mobility   pa>erns  co-­‐related?   2
  • 4. What is a community? •  A  community  is  a  group  of  people  who  have   stronger  intra-­‐group  connec5ons  than  inter-­‐group   connec5ons   3
  • 5. So we asked: •  Do  administra5ve  boundaries  based  on  history  and   geography  reflect  actual  communi5es  in  2014?   •  How  can  mobile  network  big  data  be  used  to   iden5fy  these  actual  communi5es?   4
  • 6. What are the administrative boundaries in Sri Lanka? •  9  Provinces   •  25  Districts   •  331  Divisional   Secretariat   Divisions  (DSD)   •  14021  Grama   Niladhari  Divisions   (GND)     5
  • 7. What are the administrative boundaries in Sri Lanka? •  9  Provinces   •  25  Districts   •  331  Divisional   Secretariat   Divisions  (DSD)   •  14021  Grama   Niladhari  Divisions   (GND)     6
  • 8. What are the administrative boundaries in Sri Lanka? •  9  Provinces   •  25  Districts   •  331  Divisional   Secretariat   Divisions  (DSD)   •  14021  Grama   Niladhari  Divisions   (GND)   7
  • 9. What are the administrative boundaries in Sri Lanka? •  9  Provinces   •  25  Districts   •  331  Divisional   Secretariat   Divisions  (DSD)   •  14021  Grama   Niladhari  Divisions   (GND)     8
  • 10. Which DSD talks to which DSD? •  Each  link  represents   the  raw  number  of   outgoing  and   incoming  calls   between  two  DSDs         No. of calls Low High 9
  • 11. A different picture emerges when call volume is normalized by population           ●  Strongly  connected   components  are  visible       No. of calls Low High 10
  • 12. Identifying communities 11
  • 13. Identifying communities: methodology •  The  social  network  is  segregated  such  that   overlapping  connec5ons  between   communi5es  are  minimized.   •  Strength  of  a  community  is  determined  by   modularity   •  Modularity  Q              =      (edges  inside  the  community)  –          (expected  number  of  edges  inside  the  community)             M.  E.  J.-­‐Newman,  Michele-­‐Girvan,  “Finding  and  evalua5ng  community  structure  in  networks”,  Physical  Review  E,  APS,  Vol.  69,  No.  2,  p.  1-­‐16,  204.       12
  • 14. Actual communities found by the algorithm ●  The  communi5es  we  found  are   centered  around  geographic   neighbourhoods.     ●  For  Sri  Lanka,  the  op5mal   number  of  communi5es   discovered  by  the  algorithm   was  11       13
  • 15. Actual communities found by the algorithm ●  Southern  and  Northern   provinces  have  the  highest   similarity  to  their  respec5ve   provincial  boundaries.     ●  Colombo  district  is  clustered  as   a  single  community  and   Gampaha  is  merged  with  North   Western  Province   14
  • 16. Actual communities found by the algorithm ●  Southern  and  Northern   provinces  have  the  highest   similarity  to  their  respec5ve   provincial  boundaries.     ●  Colombo  district  is  clustered  as   a  single  community  and   Gampaha  is  merged  with  North   Western  Province   15
  • 17. Zooming into a community ●  Community detection is done within identified communities ●  These communities are less less related to DSD boundaries in Colombo district 16
  • 18. How communities are connected     •  Community  consis5ng  of  Northern  province   had  the  least  connectedness  with  other   communi5es.   •  Community  consis5ng  of  Ratnapura  district   and  a  part  of  Kegalle  is  the  most  connected   community.   17
  • 19. How communities are connected with each other Region Connectedness (%) 1. Ratnapura, South Kegalle, North Eastern Colombo 74 2.Badulla, Monaragala 71 3. Anuradhapura, East Kurunegala, North Matale 71 4.Kalutara 70 5. Gampaha, Puttalam, West Kurunegala 65 6.Trincomalee 65 7.Kandy, Nuwara Eliya, North Kegalle, South Matale 63 8.Colombo without North Eastern part belongs that now belongs to 1 63 9. Galle, Matara, Hambantota 62 10. Batticaloa, Polonnaruwa, Ampara 60 11. Jaffna, Kilinochchi, Mannar, Mullaitivu 56 11 10 9 7 8 6 1 2 3 4 5 connectednesswithothercommunities Low High 18
  • 20. Future work •  Explore  the  rela5onship  between  social   network  communi5es  and  their  mobility   profiles   •  Understand  the  distribu5on  of  socio-­‐ economic  status  within  communi5es   19

×