Data Signature-Based Multiyear Analysis of BLK-NB 6-Minute Interval Datasets of 2006 to 2009

摘要

In this paper, we present a multiyear analysis of the North Luzon Expressway Balintawak-NorthBound (NLEX BLKNB) Datasets of 2006 to 2009 using the 6-minute Interval traffic volume obtained from the National Center of Transportation Studies (NCTS). Previous works have only explored the behavior and characteristics of hourly interval datasets where in, however descriptive, we think is still limited. The analysis includes extracting valuable information like common problems faced by motorist and traffic enforcers based from the trends and behaviors of the traffic flow given the traffic volume of the Balintawak corridor in NLEX for each 6-minute interval in every week of the years 2006 to 2009. By getting the data signature and applying X-Means clustering algorithm, Vector Fusion and nMDS Algorithms to obtain visualizations needed for analysis, we are able to detect 12 outliers, some of which are previously known and unknown. We are also able to extract 90 different types of potential outliers from the datasets and analyze their behavior and commonality. Moreover, we are able to realize recurring patterns and relate gathered results to different factors such as holidays and weather disturbances. In addition, we are able to gather data about what may be the role of the SBPy component in characterizing the dataset. Finally, we conclude that the 6-minute Interval dataset has provided us with an effective visualization model used for traffic data analysis and a more in-depth level of characterization of the traffic behavior in NLEX, BLK-NB.