Tracking daily mobilities: GPS based bicycle data collection, processing, and analysis snapshots

911 views
806 views

Published on

Introduction by Organizers

Seraphim Alvanides1, Godwin Yeboah1, Stefan Van der Spek2, Nico de Weghe3

1Northumbria University, UK; 2TU-Delft, Netherlands; 3Ghent University, Belgium

Topic: "Tracking daily mobilities: GPS based bicycle data collection, processing, and analysis snapshots"

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
911
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
17
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Tracking daily mobilities: GPS based bicycle data collection, processing, and analysis snapshots

  1. 1. Cycling Data Challenge Workshop - CDC2013Pre-Workshop of 16th AGILE Conference 2013Leuven – Belgium.Tuesday 14th May 2013“Bisschopskamer” room at Faculty ClubAlvanides1, Yeboah1, Van der Spek2, de Weghe3Northumbria University1; TU Delft2; Ghent University3WELCOME
  2. 2. Cycling Data Challenge Workshop - CDC2013Pre-Workshop of 16th AGILE Conference 2013Alvanides1, Yeboah2, Van der Spek3, de Weghe4Northumbria University1,2; TU Delft3; Ghent University4INTRODUCTIONTRACKING DAILY MOBILITIES: GPS BASED BICYCLE DATACOLLECTION, PROCESSING, AND ANALYSIS SNAPSHOTS
  3. 3. Overview House keeping Brief background of project Data collection and sample characteristics Challenges in data collection Challenges in data processing Remarks and the rest of the programme3Yeboah & Alvanides, Northumbria University
  4. 4. House keeping4 Internet (see paper in circulation) Exits Fire alarm Where to go for coffee Where to go for lunch Gents/Ladies
  5. 5. Aim of presentation5 To provide evidence on methods used for data collection,processing, and some analysis To share challenges faced during the data collection andprocessing phase To set the scene for subsequent presentations
  6. 6. Strands: Suggestions and demands fromliterature (Why Cycling?) There is demand for sustainable ways of living due to traffic congestion, population growth, climate change, lowphysical activity, health related issues (e.g., obesity & non-communicable diseases), sedentary lifestyles etc. Cycling as active transport one of the solutions to sustainable ways of living Calls for research to focus on understanding cycling through: investigation and knowledge discovery of cyclist’s perceptionand actual route choice experiences and preferences integrated research methods which recent technologicaladvancements may permit (e.g. GPS+GIS+GISc+ABMS)6Yeboah & Alvanides, Northumbria University
  7. 7. Why primary data collection? Secondary data is aggregated or not detailedenough (e.g. census data; surveys; more recently DfT) Lack of “detailed quality data” limits this research. To make available new scientific data on actual andrevealed route choice preferences of utility cyclistswithin the research area; not existing previously. To enable further research towards understandingconstraints and enablers for cycling; especially inrelation to transport and (indirectly) “well-being”.7Yeboah & Alvanides, Northumbria University
  8. 8. Choosing study area:Analysing UK Census 2001 & 2011801020304050607080901000 20 40 60 80 100Cumulative%ofbase(total)activity(NEEngland2011Censusasbase)Cumulative % of activity(Travel to Work by Bike across NE England )Lorenz Curve for Travel to Work by Bike – Census 2011Travel to work by BikeIndex of Dissimilarity (IoD)= 11Note: Census 2001 IoD = 5North TynesideNewcastle upon TyneSouth TynesideRest of North EastGatesheadSunderland
  9. 9. Choosing study area:Analysing Tyne & Wear Household Travel Survey9From 2003 to 2011
  10. 10. Data collection / methodological issues/ Further workGodwin Yeboah, Northumbria UniversitySTUDY AREAArea:in & aroundNewcastle uponTyneBackground map: Google Maps 2012HOMEWORK/SCHOOLSTUDY AREALEGENDOverviewSlide 10Yeboah & Alvanides, Northumbria University
  11. 11. Fieldwork planning11Extensive piloting ofsurvey instrumentswith 7 participantsEvaluated 4 GPSdevices: i-gotU GT-600;Atmel BTT08; CanmoreGT-750 (L); and QstarzBT-Q1000XT (selected)ScreeningData processing&further analysisStepwise flow(main survey)Stepwise flow(during testing)RecruitmentData collectionPlanning andPreparationInvitationYeboah & Alvanides, Northumbria University
  12. 12. Tracked sample size This work (Northumbria project within Tyneside conurbation): One wave: October-November 2011 118 initially agreed to participate In the end: 81 participants out of 111utility cyclists 79 used in this presentationLessons learnt from other related work such as: UK National Travel Survey (NTS) GPS Feasibility study (DfT) The fieldwork was done in two waves; 66 adults in one wave (October-November) and 68 adults in the second wave (January-March). In all 96 adultswere interviewed face-to-face across the two waves for the NTS study. TU Deft project in the town of Almere 15 families initially agreed to participate. However, in the end, 40 participantsout of 13 families from three neighbourhoods participated in the study bycarrying GPS devices for one week.12Yeboah & Alvanides, Northumbria University
  13. 13. Space-Time-Cube (STC) based GPS dataprocessing workflow13Yeboah & Alvanides, Northumbria University
  14. 14. Example of visual inspection:GPS raw data (left) & processed data (right)14Visualinspectionof GPS rawdataProcessed/ refineddata
  15. 15. Space-Time-Cube applicability/usability cycle15GAPYeboah & Alvanides, Northumbria University
  16. 16. Gender against number of cycle trips anddistance (km) travelled16Gender No. Over one week period per personFemale distance value is weighted to control for genderTRIPS KM(weighted)AverageKM / TRIPAverageKM /PERSONMIN / MAX(trip)Female 27 319 2137.4 6.7 79.2 0.25 km /13 kmMale 52 622 3373.0 5.4 64.9 0.12 km /36 kmTotal 79 941 5510.4 5.9 69.8
  17. 17. Trips, gender & annual household income1731%9%19%15%46%14%45%21%77%23%65%35%0%10%20%30%40%50%60%70%80%90%High Income(Distance)Low Income(Distance)High Income(Trip)Low Income(Trip)Female (f) Male (m) All (f+m)
  18. 18. Cycle trips share per employment status1859%7%16%9% 10%0%10%20%30%40%50%60%70%Participants cycle trips (%)
  19. 19. Reported travel mode by participants - t. diary1943%29%1%5%2%20%1%0%5%10%15%20%25%30%35%40%45%Bike Walk Taxi Train Bus Car OtherNumberofTrips(%)(100%=2432)Travel mode by Participants (Travel Diary)Trip (%)
  20. 20. Challenges in data collection20 Planning considerations device procurement timing, size, cost, customer support Sample, survey response, spatial distribution of trajectories Device features Battery life and the means to charge/re-charge Accuracy Memory for storing logged points Fix time. The faster the better. Mostly <=35 seconds Software for GPS device
  21. 21. GPS Logged Points212 378764116231324808 34 20 11 15Points
  22. 22. Challenges in data processing22 Non-algorithmic approach Space Time Cube usage is limited; Travel diary needed Convenient for small to medium datasets Algorithmic approach Quality assessments how reliable is the data without extra information? Non-availability of generic algorithmic tools Tool 1: Must know Java + MATSim + Eclipse http://sourceforge.net/projects/posdap/ Tool 2: Must know Java + need to conform to Copenhagen study https://github.com/bsnizek/JMapMatching
  23. 23. Our case: Network route generation23Papinski, D. & D. M. Scott (2011) A GIS-based toolkit for route choice analysis. Journal of Transport Geography, 19, 434-442.
  24. 24. Our case: An example of generatedHome-to-Work Network constrained routes24
  25. 25. Remarks and the rest of the programme Res. design: implemented in few published cycling studies No significant differences between gender and use ofcycling “corridors” Reasonable use of current cycling network (more than halfof trips take place within 20m buffer around cyclingpaths). Network data from Newcastle City Council used. However, need to improve cycling network for the 1/3 oftrips taking place “off” the network => Policy implications25Yeboah & Alvanides, Northumbria University
  26. 26. Rest of the programme Let’s go through the workshop programme Possible discussions during breaks or sessions Keynote presentations Methods and findings arising from presenters’ presentation Your reasons for attending the workshop New ideas emanating from discussions Organizers intend to take pictures during the presentationsand discussions.26Yeboah & Alvanides, Northumbria University
  27. 27. MOVE-COST:Funded CDC2013 WorkshopCHOROCHRONOS:Provided secure platform for the bike data managementAGILE2013 TEAM:Accepted and facilitated this workshopALL CONTRIBUTORS:Organizers, presenters, attendees27Yeboah & Alvanides, Northumbria UniversityPlease keep questions for the morning open discussionAcknowledgements
  28. 28. Other information:About presenter and supervision teamPhD Student:• Blog: http://godwinyeboah.blogspot.com/• YouTube Channel: http://www.youtube.com/SpatialScience• Twitter: http://twitter.com/#!/godwinyeboahSupervision team:• Dr. Seraphim Alvanideshttp://www.northumbria.ac.uk/sd/academic/bne/study/aec/acestaff/seraphimalvanides• Dr. Emine Mine Thompsonhttp://www.northumbria.ac.uk/sd/academic/bne/study/aec/acestaff/eminethompson28Yeboah & Alvanides, Northumbria University

×