Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/163118
Title: Mining big spatial data
Authors: Lee, Ming Da 
Keywords: Engineering::Computer science and engineering
Issue Date: 2022
Publisher: Nanyang Technological University
Source: Lee, M. D. (2022). Mining big spatial data. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/163118
Project: SCSE21-0647
Abstract: With the expansion and advancement of road infrastructures, many large cities are experiencing increasingly severe road traffic accidents, resulting in costly socioeconomics losses. Identification of road traffic accident hotspots enables the authority as well as road users to take preventive measures to minimize the possibility of road traffic accidents. In addition, the identification of traffic accident hotspots allows future road planners to draw insights from the data such that better planning of road infrastructures such as traffic lights, streetlamps, and speed cameras can be implemented to improve the safety of the road network. At present, this is achieved by performing spatial study to determine road traffic accident patterns followed by recognizing road location with high tendency of accidents. Even though spatial analysis is able to pinpoint road accident hotspot effectively, sole dependency on spatial analysis technique is insufficient when the desired analysis requires much larger data sets (i.e., comprising weather, time of the day, seasons, road wetness when the accident occurs) to enhance the accuracy of analyze results. To bridge the current research gap, this project proposes to adopt spatio-temporal analysis that is able to visualize enormous amount of road traffic accident data using pattern analysis techniques. This project will be using four different visualization techniques, such as hex bin, kernel density estimation, heatmap and datashader, and three different point pattern analysis techniques, namely quadrant statistic, average nearest neighbor and centrography, to analyze road traffic accident data. The road traffic accident data used in this investigation is the UK traffic accidents from year 2005 to year 2014 excluding year 2008. Subsequently, a comparison study has been carried out to investigate the performance of DBSCAN and K-mean to identify road traffic accident hotspots, followed by spatial regression analysis. Lastly, a conclusion was made to identify the superior process which aid in identifying road traffic accident hotspots.
URI: https://hdl.handle.net/10356/163118
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
fyp final report (3).pdf
  Restricted Access
2.96 MBAdobe PDFView/Open

Page view(s)

15
Updated on Nov 30, 2022

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.