A Markov chain dynamic model for trip generation and distribution based on CDR

Abstract

This study focuses on exploiting Call Detail Records (CDR) data in order to detect the demand distribution among different zones within the day, together with information about the type of activity that characterises each zone. Traffic zones are first identified and shaped through a k-means clustering analysis. Then, the traffic between different clusters is analysed with the aim to identify the type of zone by evaluating the traffic calls density for each time period. To evaluate the propagation of the demand among the different zones, Markov Chains theory is used in order to evaluate the transition matrix among different time steps. The developed model enables one to predict CDR variations in time and space and, hence, being a proxy for the trip distribution. Results point out how these matrices are similar for consecutive time intervals; therefore, it is possible to aggregate them in an hourly transition matrix, losing a small amount of information over the structure of the demand. This also shows that trip distribution varies relatively slowly in time and is spatially consistent along the days.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.