ETC Conference Papers

Seminar
Day 3 (11 Sep 2020), Session 8, Data Fusion/Integration, 13:00 - 15:00

Status
Accepted, documents submitted

Submitted by / Abstract owner
Gulcin Dalkic-Melek

Authors
Ali Afshar Kazemi
Department of Civil Engineering, METU, Ankara, Turkey,
E-mail: ali.kazemi.afshar@gmail.com

Mehmet Kurhan
Parabol Software Company, METU Technopark
mehmet.kurhan@paraboly.com

Zuhal Nalcakar
ISSD Company, METU Technopark
E-mail: zuhal.nalcakar@issd.com.tr

Assoc. Prof. Dr. Hediye TUYDES-YAMAN (Corresponding author)
Department of Civil Engineering, METU, Ankara, Turkey,
Phone: +903122107471,
Fax: +903122105401,
E-mail: htuydes@metu.edu.tr

Short abstract
This study aims to present an evaluation of the smartcard and GPS track data quality and lessons learnt during this process for the case study of Konya, Turkey.

Abstract
Public transit (PT) is a necessary but very challenging component of sustainable urban transportation, especially bus PT with very large number of lines and bus stops to manage. Increasingly deployed Intelligent Transportation Systems (ITS) tools such as automated fare collection systems (i.e., smartcard systems) has not only encouraged the use of PT with simplified transfer and flexible pricing options, but it also provided large disaggregate dataset series (recording every trip for every smartcard) over long periods. The smartcard use has been increasing in most of the metropolitan cities in Turkey, where motorization rates (especially automobile ownership) are comparatively lower than European levels, despite much larger city sizes. The availability of PT smartcard data would help to plan and manage PT services in Turkish cities, if the data can be associated properly with the urban transport infrastructure (road network, stop locations, etc.), land use (Central Business District –CBD, work places, etc.) and socioeconomic data (income, age, etc.). However, the evaluation of smartcard data from city of Konya, Turkey, showed that there are major challenges in the collection and pre-processing of data before conducting any PT evaluation (accessibility or performance analysis). These problems constitute a hidden bottleneck in the development of the advance public transit systems in developing countries, where PT data quality and standards are not defined at all.

This study aims to present an evaluation of the smartcard and GPS track data quality and lessons learnt during this process which is a part of Newton-Katip Celebi project funded by Royal Academy of Engineering, UK. In the scope of this study, first, overall data continuity in smartcard data was checked in terms of availability of data for each line, each vehicle, each day, etc. Secondly, missing data was detected for smartcard data in terms of missing cells in the row data (ie. Smartcard transaction time, bus line ID, driver ID, etc.). Then, data consistency was evaluated in order to detect inconsistencies in stop sequence, declared bus line route versus bus GPS track data. Results showed that there is 4% missing data in 6-month smartcard transaction which has around 46 million data in total. Before doing data consistency analyses, bus lines serving in urban area (160 bus lines) and stop locations were digitalized based on declared routes by municipality. The major problems were detected as scatteredness of smartcard transaction points around roads and stations, data continuity problems regarding bus GPS data at certain times and detection of off-route GPS track data. In conclusion, the data quality problems generally raised from the technological/infrastructural problems, human errors or operational instantaneous changes in schedule, etc.

Programme committee
Data

Topic
The future of transport