Time Accuracy Analysis of Post-Mediation Packet-Switched Charging Data Records for Urban Mobility Applications

TL;DR

This paper investigates the time accuracy issues of post-mediation packet-switched charging data records used in urban mobility, proposing a methodology to analyze timestamp errors to improve data reliability.

Contribution

It provides a detailed analysis of timestamp errors in post-mediation CDR and introduces a methodology for error time series analysis in network cells.

Findings

Identifies key sources of timestamp inaccuracies in post-mediation CDRs.

Proposes a systematic approach to analyze and quantify timestamp errors.

Highlights the trade-off between data accessibility and time accuracy in urban mobility applications.

Abstract

Telecommunication data is being used increasingly in urban mobility applications around the world. Despite its ubiquity and usefulness, technical difficulties arise when using Packet-Switched Charging Data Records (CDR), since its main purpose was not intended for this kind of applications. Due to its particular nature, a trade-off must be considered between accessibility and time accuracy when using this data. On the one hand, to obtain highly accurate timestamps, huge amounts of network-level CDR must be extracted and stored. This task is very difficult and expensive since highly critical network node applications can be compromised in the data extraction and storage. On the other hand, post-mediation CDR can be easily accessed since no network node application is involved in its analysis. The pay-off is in the lower accurate timestamps recorded, since several aggregations and filtering is performed in previous steps of the charging pipelines. In this work, a detailed description of the timestamp error problem using post-mediation CDR is presented, together with a methodology to analyze error time series collected in each network cell.