Identifying automatic vehicle location (AVL) data completeness issues in a rural transit authority system

Abstract

We analyzed AVL stop level data from a rural transit system to identify data completeness and systematic data capture failures. Systematic data loss could compromise the validity of further analyses of the data, such as schedule adherence or run time performance. We audited the data to identify missing values and possible data recording errors. The frequency of missing values was analyzed as a function of trip start time, stop number, day of the week, and last reported seconds late. We also perform an outlier and extreme value analysis as a function of missing records per trip. We conclude that there are systematic data capture errors in the system that needs to be addressed before further studies, such as run time analysis can be performed. Given the widespread adoption of an AVL system by rural transit system, it is recommended that detail data completeness analysis becomes routine before using the data generated to perform other studies.