All Issue

2025 Vol.43, Issue 1 Preview Page

Article

Time Series Cluster Analysis of Seoul Subway Usage Patterns under COVID-19 Based on Dynamic Time Warping

동적 시간 왜곡 기반 코로나19에 따른 서울 지하철 이용 패턴의 시계열 군집 분석

Junghee KIM1
,
Hyeyeong LEE2
,
Seyun KIM3
,
Yoonjin YOON4*
김 정희1
,
이 혜영2
,
김 세윤3
,
윤 윤진4*
1Master of Science, Department of Civil and Environmental Engineering, KAIST, Daejeon 34141, Korea
2Assistant, Kakao Mobility, Gyeongi-do 13529, Korea
3Research Professor, Department of Civil and Environmental Engineering, KAIST, Daejeon 34141, Korea
4Professor, Department of Civil and Environmental Engineering, KAIST, Daejeon 34141, Korea
1한국과학기술원 건설및환경공학과 석사
2카카오모빌리티 어시스턴트
3한국과학기술원 건설및환경공학과 연구교수
4한국과학기술원 건설및환경공학과 교수
*Corresponding Author
28 February 2025. pp. 15-37
PDF XML Citation
Abstract

For three years since early 2020, policies restricting physical movement, such as social distancing measures to prevent the spread of COVID-19 (Coronavirus Disease 2019), were implemented. This has led to a significant decrease in overall mobility, including a notable decline in subway usage, which is vital for urban transportation. However, the extent of this decrease and the patterns of change varied across different areas of the city. Identifying regions with similar mobility patterns is essential for resilient urban management. This study analyzes ridership reduction patterns at subway stations in Seoul during the pandemic using Dynamic Time Warping and Agglomerative Hierarchical Clustering. This research differs from previous studies in that it analyzes mobility changes in time series based on hourly unit data from a long-term perspective, covering the period from the outbreak of COVID-19 to the period after the complete lifting of social distancing and it defines the detailed phases of the analysis period. The results indicate that most stations did not recover to the 2019 level of daily mobility. Additionally, the groups of subway stations with relatively lower ridership levels recovered earlier than the group with higher ridership levels. As the COVID-19 pandemic progressed toward its later stages, the variation in the rate of decline in the number of users became smaller, corresponding to the social distancing policies. These findings can be used to support policy measures tailored to regional needs during future pandemics or similar shocks, ensuring efficient allocation of limited resources by prioritizing subway stations for policy application.

Keywords
agglomerative hierarchical clustering
COVID-19
dynamic time warping
subway ridership
time series analysis

2020년 초부터 3년간, 코로나19(코로나바이러스감염증-19)의 대유행으로 인해 감염 확산을 막기 위한 사회적 거리두기 등 물리적 이동을 제한하는 정책이 시행되었다. 이는 많은 사람의 통행량 감소를 초래했으며, 도시 내 핵심적인 대중교통수단인 지하철 이용량도 큰 폭으로 감소시켰다. 하지만 팬데믹(Pandemic) 기간 중 이러한 감소 폭 및 패턴은 도시 내 전 지역에서 동일하게 나타나지 않으며, 유사한 패턴을 보이는 지역군을 식별한다면 회복력 있는 도시관리에 유용하게 활용될 수 있다. 이에 본 연구는 팬데믹 동안 서울 지하철 역별 이용량 감소 패턴을 동적 시간 왜곡(Dynamic Time Warping)과 응집형 계층적 군집화(Agglomerative Hierarchical Clustering)를 통해 분석하였다. 본 연구는 서울 지하철 이용자 수를 기준으로 코로나19 발생 시점(2020.01)부터 사회적 거리두기 전면 해제 이후의 시기(2023.04)까지 장기적인 관점에서, 분석 기간의 세부 단계를 정의하고, 그룹별, 단계별로 사람들의 이동성 변화를 시간 단위 이용량 정보에 기반하여 시계열로 분석한다는 점에서 기존 연구들과의 차별성이 존재한다. 분석 결과, 대부분의 역에서 2019년도 수준의 일상적인 이동성을 회복하지 못하였고, 이용자 수 규모가 상대적으로 적은 지하철 역 그룹이 이용자 수 규모가 큰 그룹보다 조기에 회복되었으며, 코로나19 유행 후반부로 갈수록 거리두기 정책과 비례하는 이용자 수 감소율의 변동 폭이 줄어들었다. 본 연구를 통해 향후 코로나19와 같은 유행병이나 충격이 발생할 경우, 지역별 필요에 부합하는 정책 지원에 활용될 수 있을 것으로 기대하며, 정책 적용을 위한 지하철 역의 우선순위를 선정함으로써 한정된 자원을 효율적으로 활용할 수 있을 것으로 기대한다.

키워드
응집형 계층적 군집화(AHC)
코로나19
동적 시간 왜곡(DTW)
지하철 이용자 수
시계열 분석
References
1

AlMahamid F., Grolinger K. (2022), Agglomerative Hierarchical Clustering With Dynamic Time Warping for Household Load Curve Clustering, 2022 IEEE CCECE, IEEE, 241-247.

10.1109/CCECE49351.2022.9918481
2

Bellman R. (1961), A Mathematical Formulation of Variational Processes of Adaptive Type, Berkeley Symp. on Math. Statist. and Prob., University of California Press, 37.

3

Bhin M., Son S., Joh C. H. (2021), An Analysis of Spatial Characteristics of Change in The Number of Passengers at Individual Bus Stops Under COVID-19, J. Korean Soc. Transp., 39(4), Korean Society of Transportation, 447-463.

10.7470/jkst.2021.39.4.447
4

Charu C. Aggarwal, Chandan K. Reddy (2013), Data Clustering: Algorithms and Applications, CRC Press LLC, 100-101.

10.1201/b15410
5

Cho H. R., Yun S. B., Jeong Y. J. (2020a), Seoul Transportation Changes and Strategies After COVID-19, Transp. Technol. Policy, 17(3), Korean Society of Transportation, 46-51.

6

Cho J., Shin Y., Kim S., Kim N., Oh S., Cho H. et al. (2020b), Running The COVID-19 Marathon: The Behavioral Adaptations in Mobility and Facemask Over 27 Weeks of Pandemic in Seoul, South Korea, arXiv preprint arXiv:2009.09073.

7

Choi W. G., Ryu S., Jung P. H., Kang S. (2024), Differential Dynamics of Transit Use Resilience During The COVID-19 Pandemic Using Multivariate Two-Dimensional Functional Data Analysis, IEEE Access, 12, IEEE, 8721-8743.

10.1109/ACCESS.2024.3353780
8

Fransiska H. (2021), Clustering Provinces in Indonesia Based on Daily Covid-19 Cases, J. Phys. Conf. Ser., 1863(1), IOP Publishing, 012015.

10.1088/1742-6596/1863/1/012015
9

Infectious Disease Portal, https://dportal.kdca.go.kr/pot/cv/trend/dmstc/selectMntrgSttus.do, 2023.8.31.

10

Jain A. K., Murty M. N., Flynn P. J. (1999), Data Clustering: A Review, ACM Comput. Surv., 31(3), Association for Computing Machinery, 264-323.

10.1145/331499.331504
11

Joh C. H., Bhin M., Sohn S. (2023), Analysis of The Congruity of Changes in COVID-19 and Transit Use in Gyeonggi Province, Korea, Using Dynamic Time Warping and CHAID, J. Korean Soc. Transp., 41(7), Korean Society of Transportation, 861-877.

10.7470/jkst.2023.41.7.861
12

Kaufman L., Rousseeuw P. J. (2009), Finding Groups in Data: An Introduction to Cluster Analysis, John Wiley & Sons(Hoboken, New Jersey), 88.

13

Kim E., Min J. H., Kim D. K., Moon S. (2023a), Analysis on Changes in Travel Patterns and Travel Purpose of Smartcard Users Pre and Post COVID-19 Outbreak, J. Korean Soc. Transp., 41(7), Korean Society of Transportation, 771-788.

10.7470/jkst.2023.41.7.771
14

Kim S., Jang K., Yeo J. (2023b), Non-linear Impacts of COVID-19 Pandemic on Human Mobility: Lessons From Its Variations Across Three Pandemic Waves, Sustain. Cities Soc., 97, Elsevier, 104769.

10.1016/j.scs.2023.104769
15

Korea Disease Control and Prevention Agency(KDCA) (2023), COVID-19 Waves and Their Characteristics in The Seoul Metropolitan Area (Jan 20, 2020 - Aug 31, 2022), 111-136.

16

Lee C. K. H., Leung E. K. H. (2023), Spatiotemporal Analysis of Bike-share Demand Using DTW-based Clustering and Predictive Analytics, Transp. Res. E, 180, Elsevier, 103361.

10.1016/j.tre.2023.103361
17

Liu Y., Wang X., Song C., Chen J., Shu H., Wu M. et al. (2023), Quantifying Human Mobility Resilience to The COVID-19 Pandemic: A Case Study of Beijing, China, Sustain. Cities Soc., 89, Elsevier, 104314.

10.1016/j.scs.2022.10431436438675PMC9676079
18

Malik A., Tuckfield B. (2019), Applied Unsupervised Learning With R: Uncover Hidden Relationships and Patterns With K-means Clustering, Hierarchical Clustering, and PCA, Packt Publishing, 65-69.

19

McKenzie G., Adams B. (2020), A Country Comparison of Place-based Activity Response to COVID-19 Policies, Appl. Geogr., 125, Elsevier, 102363.

10.1016/j.apgeog.2020.10236333162624PMC7604168
20

Muller M. (2007), Information Retrieval for Music and Motion, Springer(Heidelberg, Germany), 69-84.

21

Nikiforiadis A., Mitropoulos L., Kopelias P., Basbas S., Stamatiadis N., Kroustali S. (2022), Exploring Mobility Pattern Changes Between Before, During and After COVID-19 Lockdown Periods for Young Adults, Cities, 125, Elsevier, 103662.

10.1016/j.cities.2022.10366235309857PMC8923996
22

Niu C., Zhang W. (2023), Causal Effects of Mobility Intervention Policies on Intracity Flows During The COVID-19 Pandemic: The Moderating Role of Zonal Locations in The Transportation Networks, Comput. Environ. Urban Syst., 102, Elsevier, 101957.

10.1016/j.compenvurbsys.2023.10195736938101PMC10011038
23

OECD (2020), COVID-19 TRANSPORT BRIEF Re-Spacing Our Cities For Resilience, 1.

24

Office of the President of the Republic of Korea (2023), President Yoon Declares 'Full Return to Normalcy' From COVID-19, https://www.president.go.kr/newsroom/press/DPYHLWWh, 2023.5.15.

25

Radovanović A., Li J., Milanović J. V., Milosavljević N., Storchi R. (2020), Application of Agglomerative Hierarchical Clustering for Clustering of Time Series Data, Proc. 2020 IEEE PES Innov. Smart Grid Technol. Eur. ISGT-Eur, IEEE, 640-644.

10.1109/ISGT-Europe47291.2020.9248759
26

Rapid Transition Alliance (2019), Reducing Roads Can Cause Traffic to 'Evaporate', https://rapidtransition.org/stories/reducing-roads-can-cause-traffic-to-evaporate/, 2024.11.12.

27

Rousseeuw P. J. (1987), Silhouettes: A Graphical Aid to The Interpretation and Validation of Cluster Analysis, J. Comput. Appl. Math., 20, Elsevier, 53-65.

10.1016/0377-0427(87)90125-7
28

Seoul Metro (2023), Seoul Metro Information on The Number of Passengers Boarding and Alighting by Day and Time Slot, https://data.seoul.go.kr/dataList/OA-12921/F/1/datasetView.do#, 2023.9.12.

29

Seoul Metropolitan Government (2020), Daily Traffic Status (sharing ratio) Statistics by Traffic Mode in Seoul, https://data.seoul.go.kr/dataList/250/S/2/datasetView.do#, 2023.9.12.

30

Seoul Metropolitan Government (2023), The Obligation to Wear Masks on Public Transportation, Including Buses and Subways, Was Lifted Starting March 20, https://news.seoul.go.kr/welfare/archives/550790, 2024.11.12.

31

Suris F. N. A., Bakar M. A. A., Ariff N. M., Mohd Nadzir M. S., Ibrahim K. (2022), Malaysia PM10 Air Quality Time Series Clustering Based on Dynamic Time Warping, Atmosphere, 13(4), MDPI, 503.

10.3390/atmos13040503
32

Wang J., Huang J., Yang H., Levinson D. (2022), Resilience and Recovery of Public Transport Use During COVID-19, npj Urban Sustain., 2(1), Springer NatUre in Partnership With The Royal Melbourne Institute of Technology (RMIT), 1-9.

10.1038/s42949-022-00061-1
33

Wang S., Huang X., Shen Q. (2024), Disparities in Resilience and Recovery of Ridesourcing Usage During COVID-19, J. Transp. Geogr., 114, Elsevier, 103745.

10.1016/j.jtrangeo.2023.103745
34

Welfare News (2020), Welfare News [Settlement 2020] Remote Classes Became A Reality due to COVID-19 First-ever Online School Opening in History, https://www.bokjiro.go.kr/ssis-tbu/cms/pc/news/news/6694925.html, 2024.4.11.

35

WHO (2023), WHO Director-General's Opening Remarks at The Media Briefing - 5 May 2023, https://www.who.int/news-room/speeches/item/who-director-general-s-opening-remarks-at-the-media-briefing---5-may-2023, 2023.5.8.

Information
  • Publisher :Korean Society of Transportation
  • Publisher(Ko) :대한교통학회
  • Journal Title :Journal of Korean Society of Transportation
  • Journal Title(Ko) :대한교통학회지
  • Volume : 43
  • No :1
  • Pages :15-37
  • Received Date : 2024-08-02
  • Revised Date : 2024-09-03
  • Accepted Date : 2024-12-23
  • DOI :https://doi.org/10.7470/jkst.2025.43.1.015
Journal Informaiton Journal of Korean Society of Transportation Journal of Korean Society of Transportation
Archives
: 2015 Vol.34~
  • NRF
  • KOFST
  • crossref cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close