This study presents a method to evaluate the daily operation of traditional
public transportation using multi-source data and rank transformation. In
contrast with previous studies, we focuses on dynamic indicators generated
during vehicle operation, while ignoring static indicators. This provides a
better reference value for the daily operation management of public transport
vehicles. Initially, we match on-board GPS data with network and stop
coordinates to extract arrival and departure timetable. This helps us calculate
dynamic operational metrics such as dwell time, arrival interval, and frequency
of vehicle bunching and large interval. By integrating IC card data with arrival
timetable, we can also estimate the number of people boarding at each stop and
derive passenger arrival time, waiting time, and average waiting time. Finally,
we developed a comprehensive dynamic evaluation method of public transportation
performance, covering the three dimensions: bus stops, vehicles, and routes.
This method uses K-means clustering to classify and applies rank transformation
techniques to score. At stop levels, we use principal component analysis(PCA) to
identify key influencing factors, anf apply K-means for clustering and
service-level classification. At the vehicle and route level, we perform rank
transformation on indicators such as average waiting time and vehicle bunching
frequency. Delphi method is used to determine the relative weights of each
indicator, so as to facilitate the ranking of bus routes according to the
comprehensive score. This method is applicable to the dynamic operation
indicators of 20 bus routes in Shenzhen, involving 293 vehicles and 506 stops.
The results show that this method can effectively evaluate the dynamic operation
of public transport and make contribution to daily management.