A Data-Driven Analysis of Traffic Patterns in Downtown San Jose, California

Publication Date

6-2026

Document Type

Conference Proceeding

Exhibition/Performance Dates

19-22 April 2026

Publication Title

2026 IEEE Conference on Technologies for Sustainability (SusTech)

Conference Location

Los Angeles, CA, USA

DOI

https://doi.org/10.1109/SusTech67720.2026.11536363

Abstract

Urban growth in San Jose has placed sustained pressure on its transportation infrastructure, leading to persistent congestion along key arterial routes. Although prior research has explored traffic volumes in larger metropolitan contexts, few studies have assessed long-term congestion patterns at the city scale using spatial clustering techniques. This study analyzes 15 years (2006-2021) of Average Daily Traffic (ADT) data from San Jose's public GIS portal, employing a three-stage analytical framework: (1) exploratory analysis of traffic volume distributions, (2) segmentation of roadway segments by ADT percentiles, and (3) application of the DBSCAN algorithm to identify persistent congestion spatial clusters. The results indicate that the volume of traffic is highly skewed and a small subset of arterial roads, most notably the Capitol Expressway, Tully Road, and Brokaw Road, consistently carry most of the traffic. These high-volume segments are geographically concentrated and stable over time, suggesting a structural pattern of congestion rather than transient or episodic spikes. Spatial clustering further reveals that a few dominant clusters that capture the majority of congestion hotspots, with one cluster alone covering 70 % of all high-ADT locations. These findings support a corridor-focused approach to traffic management, where targeted interventions, such as signal coordination, lane optimization, and intersection redesign, may have a greater impact than dispersed, citywide measures. The proposed approach can be applied to other cities focused on long-term traffic challenges. In future research, we intend to integrate crash data, land use information, and temporal traffic patterns to evaluate risk at the corridor level and inform datadriven planning efforts.

Keywords

Roads, Storage area networks, Urban areas, Printing, Timing, Labeling, Transportation, Vehicles, Planning, Tail

Department

Computer Science

Link to Full Text

Share

COinS