Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.779310
Title: Geospatial computing : architectures and algorithms for mapping applications
Author: Milton, Richard William
ISNI:       0000 0004 7965 0064
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
Beginning with the MapTube website (1), which was launched in 2007 for crowd-sourcing maps, this project investigates approaches to exploratory Geographic Information Systems (GIS) using web-based mapping, or 'web GIS'. Users can log in to upload their own maps and overlay different layers of GIS data sets. This work looks into the theory behind how web-based mapping systems function and whether their performance can be modelled and predicted. One of the important questions when dealing with different geospatial data sets is how they relate to one another. Internet data stores provide another source of information, which can be exploited if more generic geospatial data mining techniques are developed. The identification of similarities between thousands of maps is a GIS technique that can give structure to the overall fabric of the data, once the problems of scalability and comparisons between different geographies are solved. After running MapTube for nine years to crowd-source data, this would mark a natural progression from visualisation of individual maps to wider questions about what additional knowledge can be discovered from the data collected. In the new 'data science' age, the introduction of real-time data sets introduces a new challenge for web-based mapping applications. The mapping of real-time geospatial systems is technically challenging, but has the potential to show inter-dependencies as they emerge in the time series. Combined geospatial and temporal data mining of realtime sources can provide archives of transport and environmental data from which to accurately model the systems under investigation. By using techniques from machine learning, the models can be built directly from the real-time data stream. These models can then be used for analysis and experimentation, being derived directly from city data. This then leads to an analysis of the behaviours of the interacting systems. (1) The MapTube website: http://www.maptube.org.
Supervisor: Hudson-Smith, A. ; Steed, A. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.779310  DOI: Not available
Share: