Identifying Bus Lines and Outliers in Bus Routes of Rio de Janeiro

Abstract

Public transportation systems play a vital role in the mobility of cities and the quality of life of their residents. With the advent of GPS technology, it is now possible to collect vast amounts of data on the movements of public buses. In this study, we investigate if it is possible to use approach based on Needleman-Wunsch algorithm (NWA) for identifying bus lines from logged GPS tracks of buses based on a timetable from the year 2020 in Rio de Janeiro, Brazil. Further we discuss possible drawbacks of that approach and reasons why such approach could not be efficiently used. As alternative we introduce approach that relies solely on the order of bus stops encoded in geohash codes to identify the correct bus line of track, eliminating the inaccuracies caused by location interpolation used in NWA. Our findings suggest that approach based on NWA could not be efficiently used for labeling logged GPS bus tracks by bus lines because of interpolation and global alignment. But introduced alternative approach could reliable assign to bus tracks their bus line. Such approach is also able to filter out inaccurate bus tracks and outliers. Detecting anomalies and outliers can provide valuable insights into the transportation patterns of cities, and analyzing the data from public transportation systems can inform the development of more efficient and effective transportation policies.