Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Farah_Kamw_Dissertation_2019.pdf (30.56 MB)

ETD Abstract Container

Abstract Header

UTILIZING BIG TRAJECTORY DATA FOR URBAN VISUAL ANALYTICS AND ACCESSIBILITY STUDIES

Kamw, Farah Shleemon
http://rave.ohiolink.edu/etdc/view?acc_num=kent1555254527369841

Abstract Details

2019, PHD, Kent State University, College of Arts and Sciences / Department of Computer Science.
Massive urban trajectories of humans and vehicles, together with road network and Points of Interest (POIs), have been used in a wide range of research by transportation engineers and urban planning professionals. This has contributed to improve urban planning, transportation management, and knowledge of human dynamics. Interactive visual analytics tools allow a variety of users to conduct iterative visual studies over the big data with intuitive visual representations and convenient interactions. Typically, the visual analytics tasks should be conducted in three main phases: (1) Preprocessing and preparing raw trajectory data with cleaning, enrichment, aggregation, and transformations. (2) Developing efficient data structures and query operations to support interactive visual querying and analysis over big data. (3) Designing visual interface with effective and convenient human-computer interactions. Firstly, this dissertation develops data preprocessing tools of various trajectories, road networks, and POIs, which can be directly used by general users through a web-based system. Users can directly upload raw trajectory data, while the system automatically fetches corresponding road segments data from OpenStreetMap (OSM), extracts zip code regions, or creates grid rectangular regions to couple the raw GPS data with geographical context. The system also automatically matches the trajectories with these road segments or regions. Secondly, effective data models are designed to store and manage heterogeneous urban data in a spatial database called the Trajectory DataBase (TrajBase). The key contribution is to develop trajectories and road segments (or regions) based geo-indexing scheme for trajectory-based urban study. This scheme can support fast spatial-temporal queries and visualization, while the traditional geo-indexing scheme is mostly designed for point-based geo-data. Thirdly, based on the proposed data models and tools, visual analytics queries and functions are supported in open-source visual systems which benefit domain users and practitioners of a great variety. The data models and the preprocessing tools are integrated into an open source software called the Trajectory Analytics (TrajAnalytics), which is developed to meet a dire need from domain researchers to analyze their trajectory datasets. During the research, new techniques were also developed to enhance the precision of road segments of the extracted road network. Moreover, a new graph-based model called the Urban Structure Accessibility Graph (USAGraph) is further developed as a dual road network graph to construct an efficient data structure that can manage trajectories over graph database created from the enhanced road network. The graphs store and manage traffic information, taxi trip information and POI information of different time periods. They facilitate rapid graph-based computation of urban accessibility. Based on the proposed graph model, a new computational model and visualization system is developed that assists domain users to interactively study jointly-constrained accessible regions, street segments, and POIs. In particular, this system is built upon a new Min-Max Joint Set (MinMaxJS) model, where specifically-designed set operations not only represent the accessible regions but also compute the minimum and maximum access times to urban geographical structures (roads and POIs) inside the regions from joint constraints. In addition, a group of new visualization algorithms are introduced, such as the region drawing method based on a concave hull and a coloring scheme to enhance accessibility visualization. Finally, a new interactive visualization system is proposed for discovering Latent Accessibility Clusters (LACs) of POIs and revealing hidden urban regions with different accessibility patterns. The system is based on a newly proposed graph-based clustering method to form the LACs and a Voronoi-based drawing method to visualize the LACs as spatial regions on the map. The system is useful to analyze urban structures (e.g., POIs, regions) based on their disparities in the access to essential facilities and services (e.g., hospitals, shops, public transits) during different time intervals and under different traffic conditions.
Ye Zhao, Dr. (Advisor)
Feodor Dragan, Dr. (Committee Member)
Arden Ruttan, Dr. (Committee Member)
Xinyue Ye, Dr. (Committee Member)
Wei Li, Dr. (Committee Member)
194 p.
Computer Science
Information Visualization; Visual Analytics; Urban Accessibility; Urban Trajectory Data; Visual Queries of Trajectories

Recommended Citations

Citations

  • Kamw, F. S. (2019). UTILIZING BIG TRAJECTORY DATA FOR URBAN VISUAL ANALYTICS AND ACCESSIBILITY STUDIES [Doctoral dissertation, Kent State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=kent1555254527369841

    APA Style (7th edition)

  • Kamw, Farah. UTILIZING BIG TRAJECTORY DATA FOR URBAN VISUAL ANALYTICS AND ACCESSIBILITY STUDIES. 2019. Kent State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=kent1555254527369841.

    MLA Style (8th edition)

  • Kamw, Farah. "UTILIZING BIG TRAJECTORY DATA FOR URBAN VISUAL ANALYTICS AND ACCESSIBILITY STUDIES." Doctoral dissertation, Kent State University, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=kent1555254527369841

    Chicago Manual of Style (17th edition)

kent1555254527369841
299
© 2019, all rights reserved.
This open access ETD is published by Kent State University and OhioLINK.