Event detection method with principal component analysis based sensor placement
Topics: Cyberinfrastructure
, Geographic Information Science and Systems
, Transportation Geography
Keywords: big data, transportation, event detection
Session Type: Virtual Paper Abstract
Day: Saturday
Session Start / End Time: 2/26/2022 11:20 AM (Eastern Time (US & Canada)) - 2/26/2022 12:40 PM (Eastern Time (US & Canada))
Room: Virtual 20
Authors:
Yuqin Jiang, University of South Carolina
Andrey A. Popov, Virginia Tech
Zhenlong Li, University of South Carolina
,
,
,
,
,
,
,
Abstract
Human movements in urban areas are pulses of the city. Understanding human mobility patterns in a city is essential for urban planning, transportation, and event detection. In this study, we propose a principal component analysis-based event detection method using optimal sensor placement. To evaluate the effectiveness of this approach, we applied this method to analyze NYC taxi trip records from 2009 to 2012. We first applied this method to detect anomalies for the 4-year overall travel patterns. Secondly, we run this method to each year separately and detected the most significant events of each year. Lastly, we showed the effectiveness in using this method as a trip prediction. This method successfully found most expected anomalies in NYC such as Thanksgiving, New Year, Hurricane Sandy, and multiple parades. The contributions of this study include: (1) this method does not require chronologically ordered data, which reduced pre-processing time; (2) results in this method include RMSE at multiple spatiotemporal resolutions, which can be easily adapted to further analysis for different applications; and (3) this method not only can be used for event detections with existing historical travel data, but also can help predict future traffic patterns.
Event detection method with principal component analysis based sensor placement
Category
Virtual Paper Abstract
Description
This abstract is part of a session. Click here to view the session.
| Slides