Exploring the Geography of Access to Fixed Line Broadband Services in England through Crowd Sourced Speed Check Data

Talk given at the Association of American Geographers annual conference, Los Angeles, April 2013

Transcript

1. Exploring  the  Geography  of  Access  to   Fixed  Line  Broadband  Services  in   England  through  Crowd  Sourced  Speed   Check  Data   Dean  Riddlesden     Department  of  Geography  and  Planning     University  of  Liverpool  
2. Introduc<on  to  the  Research  Project     •  Access  to  high  speed  broadband  has  become  increasingly  important  in   recent  years   •  Universal  Service  Commitment  (introduced  2010)  aimed  to  deliver  a   minimum  connec<on  speed  of  2  Mbps  to  every  household  in  the  UK  by   2012     •  Government  are  keen  to  get  everybody  online  –  jus<fied  by  key  economic   advantages  and  increased  compe<<veness         •  Outside  of  government  and  Ofcom  reports  there  have  been  few  academic   projects  that  aNempt  to  analyse  the  socio-­‐spa<al  structure  of  broadband   access     •  There  have  been  none  that  use  crowd-­‐sourced  speed  test  data  as  a  proxy   for  access  and  speed  
3. The  Data   •  Crowd-­‐sourced  dataset  of  around   2.8m  Internet  speed  test  results   complied  through   www.broadbandspeedchecker.co.uk   •  Data  was  collected  between  1st   January  2010  to  31st  Jan  2011   •  AVer  null  values  and  outliers  were   removed,  the  dataset  stood  at   2,085,236  geo-­‐tagged  (Postcode)   speed  test  results     •  ONS  Postcode  Directory  used  to   append  postcode  aNribute  data  to   speed  test  database  
4. •  A  second  dataset  of  UK  telephone  exchange  loca<ons  was  used  to  calculate   distances  between  speed  test  loca<ons  and  the  nearest  telephone  exchange   (Performed  in  R  using  Euclidian  distance  measurement)   •  Observa<ons  for  Wales,  Scotland  and  Northern  Ireland  removed   •  Observa<ons  where  download  speed  was  below  512Kbps  (0.5Mbps)  or  over   100,000Kbps  (97.7Mbps)  were  removed     •  Observa<ons  with  exchange  distances  over  9KM  were  removed     •  END  RESULT  –  A  comprehensive  dataset  of  over  2m  speed  test  results  geo-­‐tagged   by  postcode       •  Each  has  corresponding  aNribute  informa<on  for  that  postcode,  as  well  as  the   distance  to  the  closest  exchange     Valida<on  
5. Benefits  and  Limita<ons  of  the  Dataset   •  Very  Large  (Ofcom  average  speeds  currently  based  on  profiling  of  just  1325  user   connec<ons)   •  Fixed  geographic  resolu<on  (allows  for  aggrega<ng  to  different  geographic  levels   e.g.  LSOA)     •  Some  Postcodes  have  100’s  of  test  results,  others  have  few.  Counts  vary.     •  ‘real  world’  results  –  even  those  that  are  slow  (Ofcom  do  not  profile  anything   below  2Mbps  and  anything  not  being  sold  as  ‘up  to  10  Mbps’)   •  Speed  tests  were  performed  by  an  end  user;  they  give  a  ‘snapshot’  es<mate.     •  No  weigh<ng,  no  monitoring  for  op<mal  condi<ons  etc…  (As  with  previous  studies   and  Ofcom  techniques)    
6. Profiling  and  Results   •  Geographic  Distribu<on  of  Speed  Tests   •  Distribu<on  of  speed  tests  by  distance  to  the  closest  exchange     •  Average  Speeds  and  Distance  to  the  Nearest  Exchange     •  Speed  and  Indicators  of  Socio-­‐Spa<al  structure   •  Speed  and  Indicators  of  Rurality     •  Speed  and  Indicators  of  Depriva<on     •  A  Snapshot  of  Average  Download  Speeds  by  English  District  
7. under 3896 3896 to 5970 5970 to 8634 8634 to 12768 12768 to 20652 over 20652 Frequency of Speed Tests by English District
8. under 0.03 0.03 to 0.04 0.04 to 0.05 0.05 to 0.06 0.06 to 0.2 over 0.2 Speed Tests Per Head of Population by English District
9. Distribu<on  of  speed  tests  by  distance  to   the  closest  exchange       0 50000 100000 150000 0 2500 5000 7500 Distance to Exchange (m) Count
10. Average  Speeds  and  Distance  to  the   Nearest  Exchange       0 2000 4000 6000 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 7500 + Distance From Exchange (Metres) Mean Download Speed (Kb/s)
11. Speed  and  Indicators  of  Socio-­‐Spa<al   structure     0 2000 4000 6000 1A: Terraced Blue Collar 1B: Younger Blue Collar 1C: Older Blue Collar 2A: Transient Communities 2B: Settled in the City 3A: Village Life 3B: Agricultural 3C: Accessible Countryside 4A: Prospering Younger Families 4B: Prospering Older Families 4C: Prospering Semis 4D: Thriving Suburbs 5A: Senior Communities 5B: Older Workers 5C: Public Housing 6A: Settled Households 6B: Least Divergent 6C: Young Families in Terraced Homes 6D: Aspiring Households 7A: Asian Communities 7B: Afro−Caribbean Communities Output Area Classification Mean Download Speed (Kb/s)
12. OAC  Test  Rates   0.0 0.5 1.0 1.5 2.0 1A: Terraced Blue Collar 1B: Younger Blue Collar 1C: Older Blue Collar 2A: Transient Communities 2B: Settled in the City 3A: Village Life 3B: Agricultural 3C: Accessible Countryside 4A: Prospering Younger Families 4B: Prospering Older Families 4C: Prospering Semis 4D: Thriving Suburbs 5A: Senior Communities 5B: Older Workers 5C: Public Housing 6A: Settled Households 6B: Least Divergent 6C: Young Families in Terraced Homes 6D: Aspiring Households 7A: Asian Communities 7B: Afro−Caribbean Communities Output Area Classification Test Rate Per Current Postcode
13. Speed  and  Indicators  of  Rurality   0 2000 4000 Urban Greater than/Equal to 10K − Sparse Town and Fringe − Sparse Village − Sparse Hamlet and Isolated Dwelling − Sparse Urban Greater than/ Equal to 10K − Less Sparse Town and Fringe − Less Sparse Village − Less Sparse Hamlet and Isolated Dwelling − Less Sparse Urban/Rural Indicator Mean Download Speed (Kb/s)
14. Data  Distribu<on  –  Urban/  rural  Index   Urban/ Rural Index Count % of Total Urban >= 10K Sparse 4299 0.21 Town & Fringe Sparse 9131 0.44 Village Sparse 12774 0.61 Hamlet & Isolated Dwelling Sparse 7553 0.36 Urban >= 10K Less Sparse 1546732 74.18 Town & Fringe Less Sparse 209783 10.06 Village Less Sparse 207818 9.97 Hamlet and Isolated Dwelling Less Sparse 87146 4.18 !
15. Speed  and  Indicators  of  Depriva<on       0 2000 4000 6000 Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 IMD Decile Mean Download Speed (Kb/s)
16. A  Snapshot  of  Average  Download  Speeds  by  English   District  
17. Average  Download  Speeds  by  Hour  (Na<onal)     • • • • • • • • • • • • • • • • • • • • • • • • 4500 5000 5500 6000 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour Mean Download Speed (Kb/s)
18. Average  Download  Speeds  by  Hour  and  Rurality     2000 4000 6000 22 23 17 18 5 15 14 7 12 16 19 13 11 20 9 21 0 10 8 1 6 4 3 2 Hour Mean Download Speed (Kb/s) URBAN_RURAL_INDEX Hamlet and Isolated Dwelling Less Sparse Hamlet and Isolated Dwelling Sparse Town and Fringe Less Sparse Town and Fringe Sparse Urban >= 10K Less Sparse Urban >= 10K Sparse Village Less Sparse Village Sparse
19. Conclusions   •  It  can  be  argued  that  the  current  geography  of  access  is  in   real  terms,  near  universal     •  Profiling  the  data  by  indicators  of  rurality  confirms  that   some  less  densely  populated  and  isolated  areas  may  not  be   benefiing  from  adequate  performance     •  Access  may  be  near  universal  but  adequate  speed  is  not.     •  Prevailing  socio-­‐economic  condi<ons  of  small  geographic   areas  do  not  have  a  direct  impact  on  access  to  fixed  line   broadband  services  (IMD)    
20. Conclusions   •  Areas  of  high  depriva<on  may  oVen  have  higher   speeds  (Many  deprived  areas  are  of  high  popula<on   density  –  hence  beNer  infrastructure)     •  Poten<al  for  reaching  out  to  these  communi<es     •  There  are  rela<vely  large  fluctua<ons  in  speed   throughout  the  day  when  profiled  at  a  na<onal  level.     •  ‘Peak  Hours’  have  a  clear  impact  on  download  speeds   regardless  of  where  users  live.    
21. Speed,  Depriva<on  and  Rurality   Hamlet and Isolated Dwelling Less Sparse Hamlet and Isolated Dwelling Sparse Town and Fringe Less Sparse Town and Fringe Sparse Urban >= 10K Less Sparse Urban >= 10K Sparse Village Less Sparse Village Sparse 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 IMD Decile (1 Represents Most Deprived) Mean Download Speed (Kb/s)