Extraction of bicycle commuter trips from day-long GPS trajectories

Gerald Richter; Christian Rudloff; Anita Graser

Extraction of bicycle commuter trips from day-long GPS trajectories

Gerald Richter (Speaker), Christian Rudloff, Anita Graser

Digital Resilient Cities

Research output: Chapter in Book or Conference Proceedings › Conference Proceedings with Oral Presentation

Abstract

Collected GPS data often lacks metadata (e.g. mode of transport information) and is noisy (e.g. due to longer stops and bad GPS reception). Several measures are necessary to pre-process GPS data in order to detect trip starts and ends as well as modes of transport. These measures comprise segmentation (splitting of tracks into trips) and outlier removal. In this paper, previously developed methods implementing such algorithms and their application to the provided set of CDC2013 cyclist commuter tracks are presented. Mode detection is applied to the trips subsequently in order to separate bicycle trips from trips by other modes of transport (e.g. walking the bike). The processing results are evaluated by comparison against provided cleaned reference data. Some statistical properties of the resulting tracks are shown.

Original language	English
Title of host publication	Understanding urban cycling: A data challenge (CDC2013) - Proceedings of abstracts
Publication status	Published - 2013
Event	Understanding urban cycling: A data challenge (CDC2013) - Duration: 14 May 2013 → …

Conference

Conference	Understanding urban cycling: A data challenge (CDC2013)
Period	14/05/13 → …

Research Field

Former Research Field - Mobility Systems

Cite this

@inproceedings{ac71d1b37b564b48be7936149f9447db,

title = "Extraction of bicycle commuter trips from day-long GPS trajectories",

abstract = "Collected GPS data often lacks metadata (e.g. mode of transport information) and is noisy (e.g. due to longer stops and bad GPS reception). Several measures are necessary to pre-process GPS data in order to detect trip starts and ends as well as modes of transport. These measures comprise segmentation (splitting of tracks into trips) and outlier removal. In this paper, previously developed methods implementing such algorithms and their application to the provided set of CDC2013 cyclist commuter tracks are presented. Mode detection is applied to the trips subsequently in order to separate bicycle trips from trips by other modes of transport (e.g. walking the bike). The processing results are evaluated by comparison against provided cleaned reference data. Some statistical properties of the resulting tracks are shown.",

author = "Gerald Richter and Christian Rudloff and Anita Graser",

year = "2013",

language = "English",

booktitle = "Understanding urban cycling: A data challenge (CDC2013) - Proceedings of abstracts",

note = "Understanding urban cycling: A data challenge (CDC2013) ; Conference date: 14-05-2013",

}

TY - GEN

T1 - Extraction of bicycle commuter trips from day-long GPS trajectories

AU - Rudloff, Christian

AU - Graser, Anita

A2 - Richter, Gerald

PY - 2013

Y1 - 2013

N2 - Collected GPS data often lacks metadata (e.g. mode of transport information) and is noisy (e.g. due to longer stops and bad GPS reception). Several measures are necessary to pre-process GPS data in order to detect trip starts and ends as well as modes of transport. These measures comprise segmentation (splitting of tracks into trips) and outlier removal. In this paper, previously developed methods implementing such algorithms and their application to the provided set of CDC2013 cyclist commuter tracks are presented. Mode detection is applied to the trips subsequently in order to separate bicycle trips from trips by other modes of transport (e.g. walking the bike). The processing results are evaluated by comparison against provided cleaned reference data. Some statistical properties of the resulting tracks are shown.

AB - Collected GPS data often lacks metadata (e.g. mode of transport information) and is noisy (e.g. due to longer stops and bad GPS reception). Several measures are necessary to pre-process GPS data in order to detect trip starts and ends as well as modes of transport. These measures comprise segmentation (splitting of tracks into trips) and outlier removal. In this paper, previously developed methods implementing such algorithms and their application to the provided set of CDC2013 cyclist commuter tracks are presented. Mode detection is applied to the trips subsequently in order to separate bicycle trips from trips by other modes of transport (e.g. walking the bike). The processing results are evaluated by comparison against provided cleaned reference data. Some statistical properties of the resulting tracks are shown.

M3 - Conference Proceedings with Oral Presentation

BT - Understanding urban cycling: A data challenge (CDC2013) - Proceedings of abstracts

T2 - Understanding urban cycling: A data challenge (CDC2013)

Y2 - 14 May 2013

ER -

Extraction of bicycle commuter trips from day-long GPS trajectories

Abstract

Conference

Research Field

Fingerprint

Cite this