GISRUK 2023
University of Glasgow
Filter displayed posters (97 keywords)
Sponsor GISRUK 2023 (2)
Traffic flow (2)
geospatial (2)
show more...
Aerial (1)
Aerial images (1)
Bias (1)
Building footprints (1)
COVID-19 (1)
Classification (1)
Cluster Analysis (1)
Crime (1)
Deep Learning (1)
Deep learning (1)
Flash flood susceptibility (1)
France (1)
Fusion (1)
GIS (1)
Geodemographic (1)
Geospatial Analysis (1)
Google Earth Engine (1)
Habitat fragmentation (1)
Historical maps (1)
Human mobility (1)
Hybrid geographies (1)
In-app data (1)
K-means (1)
Lake District (1)
Land Use (1)
Land use change (1)
Leicester (1)
London (1)
Low Traffic Neighbourhood (1)
Machine Learning (1)
Multimode (1)
PGIS (1)
Participatory Mapping (1)
Pixel based analysis (1)
Remote sensing (1)
ResNet (1)
SHAP (1)
Segmentation (1)
Social Media Analysis (1)
Spatial Durbin model (1)
Spatiotemporal Data (1)
Sustainable Mobility (1)
Temporal Geodemographics (1)
Temporal-spatial (1)
Topographic map (1)
Tourism (1)
Transport (1)
Tree Preservation Order (1)
Tree-based Ensemble Learning (1)
UK census (1)
Urban areas (1)
Urban big data (1)
User Generated Content (UGC) (1)
Variables per km of route (1)
Zonal Statistics (1)
anti-immigration sentiment (1)
betting shop (1)
bicycle trajectories (1)
big data (1)
commons (1)
commuting (1)
crime attractors and generators. (1)
cycling (1)
data processing (1)
data visualisation (1)
early career (1)
gambling (1)
geodemographics (1)
in-app data (1)
inequity (1)
innovation (1)
land tenure (1)
mobility (1)
natural language processing (1)
neocartography (1)
post-disaster (1)
public transport (1)
risky facilities (1)
rural geography (1)
satellite data (1)
social disorganisation theory (1)
social media (1)
space-time cube (1)
spatial interaction (1)
start-up (1)
support (1)
survey (1)
temporal analytics (1)
traffic analysis (1)
transportation (1)
trust (1)
urban flooding; impact assessment; traffic simulation (1)
urban inequalities (1)
walking (1)
Show Posters:
Detecting bias within location histories collected from a panel of mobile applications
Hamish Gibbs, Rosalind M. Eggo, James Cheshire
Abstract
Location data collected by mobile applications is typically aggregated from a heterogeneous sample of mobile devices with varying demographic and behavioral characteristics. In this paper, we present a method for detecting homogenous groups of mobile devices relevant to specific scientific domains. We apply this method to an anonymized in-app location dataset of ~4,500,000 mobile devices in the United Kingdom, to detect homogeneous groups of devices relevant to applications including transport planning and infectious disease modeling.
Presented by
Hamish Gibbs
<Hamish.Gibbs.21@ucl.ac.uk>
Institution
Department of Geography, University College London
Other Affiliations
Department of Infectious Disease Epidemiology, London School of Hygiene & Tropical Medicine
Keywords
Human mobility, Bias, In-app data
Geovation from Ordnance Survey
Luke Hampson
Abstract
Geovation is here to maximise positive impact by bringing people together and unlocking their potential so that we have a commercially, socially and environmentally sustainable future. Geovation can support you with your next step, turning your research into a commercial proposition.
Presented by
Luke Hampson
Institution
Geovation from Ordnance Survey
Keywords
geospatial, innovation, support, start-up, Sponsor GISRUK 2023
Leicester Temporal Geodemographic Classification
Nouh AL-Mahrouqi and Stefano De Sabbata
Abstract
The city of Leicester (UK) has seen large socio-demographic changes over the past century, gaining complex social, economic and multicultural geographies. Such changes are reflected in the data collected through the census and can be illustrated by developing a temporal geodemographic classification. This paper presents the results of a Master's dissertation project aimed at developing a temporal geodemographic classification for Leicester's urban area, including the Leicester Local Authority District, as well as the Oadby and Wigston Local Authority District, analysing its changing socio-economic structure over three censuses.
Presented by
Nouh Al Mahrouqi and Dr. Stefano De Sabbata
Institution
School of Geography, Geology and the Environment, University of Leicester
Keywords
Temporal Geodemographics, Leicester, Cluster Analysis, K-means, UK census
Urban Big Data Centre
Martin Shannon
Abstract
The Urban Big Data Centre is delighted to sponsor GISRUK 2023 and warmly welcomes all delegates to Glasgow.
Presented by
UBDC
<ubdc@glasgow.ac.uk>
Institution
UBDC at the University of Glasgow
Keywords
Sponsor GISRUK 2023
Using Interactive Space-Time Cube Visualisation for Pattern Mining in Bicycle Trajectories and Traffic-Related Parameters
Andreas Keler, Sylvia Ludwig and Chenyu Zuo
Abstract
This work focuses on the visualisation of bicycle trajectories at an urban junction in Munich, Germany, using a Space-Time cube (STC) approach. The trajectories are obtained from a traffic observation using computer-vision-based approaches and pre-processed for analysis. A GUI implementation in MATLAB is introduced for evaluating the usefulness of the STC technique for transport planning and engineering purposes, with a focus on evaluating traffic safety. The visual patterns are evaluated by experts based on the quality of five interactive components of the implemented STC GUI.
Presented by
Andreas Keler
Institution
Applied Geoinformatics, University of Augsburg
Other Affiliations
2) Chair of Traffic Engineering and Control, Technical University Munich 3) Chair of Cartography and Visual Analytics, Technical University Munich
Keywords
data processing, data visualisation, space-time cube, traffic analysis, bicycle trajectories
Spatial-temporal Transformation of Habitat Causes the Northward Migration of Wild Elephants in Yunnan: Based on Google Earth Engine
Donghao Jiang and Yue Zhang
Abstract
From 2020 to 2021, a herd of wild elephants migrated more than 500 kilometers northward from their habitat (Yunnan, Southwest China) became an international hotspot. The study analysis the main habitat (Mengyang Reserve and surrounding areas) of this group of wild elephants from the natural environment, changes in land use types, and habitat fragmentation and found the rare hot and dry weather before the wild elephant migration and the habitat destruction caused by human activities in the past ten years contributed to this migration.
Presented by
Donghao Jiang
Institution
University of Manchester
Other Affiliations
Department of Geography
Keywords
Temporal-spatial, Land use change, Google Earth Engine, Habitat fragmentation
Trust in post-blast Beirut: combing survey and satellite data
Elisabetta Pietrostefani
Abstract
On the 4th of August 2020, a large amount of ammonium nitrate stored at the port of the city of Beirut exploded causing at least 200 deaths and over 7,000 injuries. This study explores the impact of the blast on trust in local institutions by analysing the degree of damage to buildings in two heavily affected neighbourhoods using satellite imagery and spatial survey data. I combine Maxmar WorldView-2 data from before and after the blast with survey data from 2018 and 2021 to evaluate the changing level of trust in governance post-disaster. The Beirut Blast had both physical and physiological channels of influence. This paper focuses on how different types of exposure to the blast changed measures of trust in local institutions. I adopt two methodological strategies, a pixel-based supervised classification of affected areas using a Random Forest classifier, and a difference-in-difference fixed effects framework to evaluate the effects of the blast on measures of trust. Results reveal that trust in institutions was already declining before the blast, and residents of heavily damaged buildings had less trust in institutions, even after accounting for socio-economic heterogeneity. Furthermore, the blast had a significant negative effect on trust in institutions. These findings highlight the need for effective disaster management policies and strategies that address the psychological and physical impact of disasters on affected communities.
Presented by
Elisabetta Pietrostefani
Institution
University of Liverpool
Other Affiliations
London School of Economics and Political Science
Keywords
trust, post-disaster, satellite data, survey, urban inequalities
Modified Point-Voxel CNN on Aerial Lidar Data
Fauzy Othman and Phil Bartie
Abstract
This research uses modified Point Voxel CNN as a method to segment point cloud data directly in 3D coordinates. In addition to xyz coordinates, geometric properties were calculated from the point cloud and added as attributes, used in training and inference. By combining voxel and multi layer perceptron methods, neighbouring point features are captured and learnt in this deep learning method. This method consumes low GPU memory size with significantly lower training time. Future work will explore further improvement in f1-score.
Presented by
Fauzy Othman
Institution
Heriot-Watt University
Other Affiliations
PETRONAS
Keywords
Deep Learning, Fusion, Aerial, Classification, Segmentation
Spatial Patterns of Transportation Mobility by Mode and Associated Neighbourhood Characteristics in Dublin, Ireland
Kevin Credit
Abstract
This paper explores the spatial patterns of walking, cycling, public transport, and automobile mobility in Dublin using a novel methodology that queries all neighbourhood-to-neighbourhood travel times within the region (by mode). These travel times are converted to travel costs and divided by Euclidean distance to create relative measures of mobility (cost-distance ratios). In terms of city-wide modal trade-offs, cycling outcompetes automobile travel through trips of 2.4km or less, while public transport outcompetes all other modes for trips longer than 8.1km. Additionally, analysing the neighbourhood characteristics of the origins and destinations for these ratios shows that neighbourhoods with more walkable built environments tend to have better observed mobility for the walking and cycling modes.
Presented by
Kevin Credit
Institution
Maynooth University
Keywords
transportation, spatial interaction, commuting, public transport, walking, cycling, inequity, mobility
Variable construction process to understand cyclists' route choice from GPS records of mobile applications
Laura Daniela Ramírez-Leuro, Lenin Alexander Bulla-Cruz
Abstract
This document summarizes the process for the treatment of GPS records of mobile applications and builds variables to understand the route decisions of cyclists. Therefore, improve public policies of sustainable mobility from a geographical analysis of road safety, infrastructure, and public safety (thefts) parameterized by km of the route traveled, which are subsequently analyzed statistically. Finally, this case study is based on the GPS records of the Biko mobile application in Bogotá, Colombia. The process can be replicated in any city with the available information.
Presented by
Laura Daniela Ramirez Leuro
Institution
Department of Urban Planning, Technische Universität Berlin
Other Affiliations
Department of Civil and Agricultural Engineering, Faculty of Engineering Universidad Nacional de Colombia
Keywords
Sustainable Mobility, Geospatial Analysis, Variables per km of route
Water telling stories Community cartographies for the survival of the amphibian culture and sustainable development in the Lower Sinú wetlands, Department of Córdoba, Colombia
Luis Miguel Sánchez Zoque & Daneris Alberto Herrera Mestra
Abstract
This article presents the experience of appropriation of the hybrid geographies approach as a support for initiatives to transform socio-environmental and territorial conflicts in the Ciénaga Grande del Bajo Sinú, in the Colombian Caribbean region, where communities are working to position their vision of the territory vis-à-vis State institutions, in order to advance in the construction of proposals that guarantee the sustainable permanence of the way of life of peasant, fishing, indigenous and Afro- descendant communities in the face of the radical transformations that the ecosystem in which they live and from which they derive their livelihoods.
Presented by
Luis Miguel Sánchez Zoque
<lmsanchezz@unal.edu.co>
Institution
Maestría en Desarrollo rural Universidad Autónoma Metropolitana, México.
Other Affiliations
Movimiento Social el Agua Contando Historias
Keywords
Hybrid geographies, neocartography,rural geography, land tenure, commons
Long-time series, small-area statistics for sustainable transport
Malcolm Morgan, and Philips I.
Abstract
Transport policy often suffers from a paucity of data on which to build a robust evidence base. Long- time series, small-area transport statistics over a large geographical area can provide insights into the effects of a broad range of transport policies. Unfortunately there are important gaps in the UK’s small-area statistics that limit their usability and effectiveness for answering key policy questions. This paper reviews the state of data in the UK and reports early progress on filling data gaps.
Presented by
Malcolm Morgan
Institution
Institute for Transport Studies, University of Leeds, UK
Keywords
Transport, Land Use, Zonal Statistics, Spatiotemporal Data, Geodemographic
Using Twitter data to analyse the spatial patterns of online anti-immigration sentiment in the UK
Matt Mason and Francisco Rowe
Abstract
There is a growing academic literature examining anti-immigration sentiment posted onto social media platforms, with evidence emerging of its impact on rises in “real physical” incidents of hate. Despite this, little is understood about the spatial pattern of the production of online anti-immigration content and contextual factors contributing to shaping its spatial configuration. This study aims to use Twitter data and natural language processing to analyse spatial patterns of online sentiment towards immigration across sub-regional areas of the UK, and identify key demographic and contextual factors associated with the production of anti-immigration sentiment on social media platforms.
Presented by
Matt Mason
Institution
Geographic Data Science Lab, The University of Liverpool
Keywords
anti-immigration sentiment, social media, natural language processing, early career
The use of in-app data to drive geodemographic classification of activity patterns
Mikaella Mavrogeni, Paul Longley, Justin van Dijk
Abstract
We use location data from multiple mobile phone applications to describe daily, weekly, seasonal and annual activity patterns. Geodemographics, or ‘the analysis of people by where they live’, provides an organising framework, extended to represent the ways in which neighbourhood residents interact with workplaces, recreational and leisure destinations and transport infrastructure. We evaluate how in-app location data can be incorporated into geodemographic analysis to better understand the flux of activity patterns that characterise densely populated areas throughout the day. Limitations and net benefits of in- app location data are critically assessed to evaluate the ways in which activity-based geodemographics are robust, effective and safe to use when characterising the population at large.
Presented by
Mikaella Mavrogeni
Institution
Department of Geography University College London
Keywords
geodemographics, big data, temporal analytics, in-app data, geospatial
Betting shop provisioning and crime patterns in England
Oluwole Adeniyi, Ferhat Tura and Andy Newton
Abstract
Critics alleged that there is a link between gambling provisioning and crime and scholars argue that gambling outlets serve as attractors of anti-social and criminal behaviours. This study compared the patterns of betting shops with different categories of crime in England. using a spatial longitudinal approaches b at neighbourhood level. Spatial analysis and multilevel negative binomial model results identified significant effect of betting shops on crime, even after accounting for ethnic heterogeneity, concentrated disadvantage, and other neighbourhood factors. Our findings suggest risk factors of crime including betting shops converge in similar areas, and this creates opportunities for more crimes.
Presented by
Oluwole Adeniyi
Institution
School of Management, Nottingham Business School, Nottingham Trent University
Other Affiliations
2) Department of Social Sciences and Social Work, Bournemouth University 3) Department of Criminology and Criminal Justice, School of Social Sciences, Nottingham Trent University
Keywords
Crime, gambling, betting shop, social disorganisation theory, risky facilities, crime attractors and generators.
Identifying Tree Preservation Order Protected Trees by Deep Learning in Greater London Area
Qiaosi Li, Mingkang Wang, Yang Cai, Qianyao Luo, Zian Wang, Qunshan Zhao
Abstract
Tree Preservation Order (TPO) is used to protect specific trees from damage and destruction, which is determined in high subjectivity. This research collected and analyzed TPO data, aerial images, geographic data, and socio-economic data in the Greater London area and developed a multi-input deep learning (DL) framework to classify TPO-protected and non-TPO-protected trees. The synergy use of aerial images and GIS data with the fusion model of ResNet50 and multilayer perceptron network produced the best classification accuracy of 87.32%. The result indicated the robustness of the multi- input DL model to identify the social attributes of trees compared with the single-input DL model.
Presented by
Qunshan Zhao
<Qunshan.Zhao@glasgow.ac.uk>
Institution
Urban Big Data Centre, School of Social and Political Sciences, University of Glasgow
Keywords
Aerial images, Deep learning, ResNet, Tree Preservation Order
The Potential of Social Media Analysis for Tourism in Indonesia
Raidah Hanifah, Alexis Comber, Nicholas Malleson and Victoria Houlden.
Abstract
Nowadays, social media are recognized as a media for information aggregation rather than just a sociability and gathering platform. Because of the exponential growth of social media usage, there has been an explosion of social media data produced by its users as UGC (User Generated Content), which can be stored in a variety of formats, including textual data, images, videos, audio, and geolocations. This potential can be used to explain various phenomena and solve problems in the tourism sector. This study will examine the use of social media data to gain a better understanding of Indonesian tourism.
Presented by
Raidah Hanifah
Institution
University of Leeds
Other Affiliations
Institut Teknologi Sumatera
Keywords
Social Media Analysis, Tourism, User Generated Content (UGC)
Finding Building Footprints in Over-detailed Topographic Maps
Robin Roussel, Ali Asadipour and Sam Jacoby
Abstract
Building footprints are a key component of many GIS applications, including morphological and street view based analysis. Crowdsourced data such as OpenStreetMap (OSM) is widespread but not consistently detailed enough to reliably extract footprints of individual buildings, while topographic building maps such as the Ordnance Survey MasterMap Topography Layer (MTL) may split footprints into multiple polygons or include constructions without an address. We propose a method to determine which topographic building polygons can be unambiguously matched to individual footprints of buildings with an address, to enable integration with address-based data sources such as transactions or energy performance certificates. The results suggest that this method recovers significantly more building footprints than what can be obtained from OSM.
Presented by
Robin Roussel
Institution
Computer Science Research Centre, Royal College of Art
Other Affiliations
School of Architecture, Royal College of Art
Keywords
Topographic map, Building footprints
Participatory Mapping?
Timna Denwood and Jonathan Huck
Abstract
This research explores the potential of developing Participatory Mapping interfaces that do not include an explicit map or other spatial contextualisation, using a case study in the Lake District, UK. Whilst there are countless examples of the successful deployment of map-based interfaces for collecting spatial information from citizens, in some cases, a map may not be an appropriate tool and other types of visual representations more suitable. We will demonstrate that the uncritical use of spatial contextualisation can impact both the usability of the interface for the user and robustness of the resultant data for the researcher.
Presented by
Timna Denwood
Institution
University of Manchester
Other Affiliations
MCGIS
Keywords
PGIS, Participatory Mapping, Lake District
Quick Mapping of historical urbanization in France, case study of Scan Histo 1950 using spatial pixel based approach
Walid Rabehi, Rémi Lemoy, Marion Le Texier
Abstract
Automatic detection of human settlement and cities forms on historical maps remains a challenging topic regarding the complexity of the maps semiology, but also because of the general conditions of maps (ink degradation, color change...). The aim of this contribution is to establish an urban layer of France based on historical maps of 1950 using pixel-based approach and multitemporal landcover map. This work will use the produced layers to test whether large cities are more or less parsimonious in terms of land use than smaller cities, at different periods.
Presented by
Walid Rabehi
<walid.rabehi@univ-rouen.fr>
Institution
UMR IDEES 6266/CNRS, University of Rouen
Other Affiliations
University of Montpellier
Keywords
Pixel based analysis, Remote sensing, Historical maps, Urban areas, France
The Influence of Low Traffic Neighbourhood Scheme on Multimodal Traffic Flow in London
Xianghui Zhang, Tao Cheng
Abstract
This study aims to investigate the influence of Low Traffic Neighbourhood (LTN) Scheme deployed since the COVID-19 pandemic on multimodal traffic flow in London. We adopt a mobile phone application dataset to investigate the changes in multimodal traffic flow generated by the general public following the introduction of LTNs. Three LTNs located in London are explored between 4th May and 30th August 2020. The analysis approved that LTN scheme could encourage residents to take cycling and restrict through-traffic, but the influence varies across areas, travel modes and groups and may affect by specific measures.
Presented by
Xianghui Zhang
Institution
SpaceTimeLab for Big Data Analytics, Department of Civil, Environmental and Geomatic Engineering, University College London
Keywords
Multimode, Traffic flow, Low Traffic Neighbourhood, London
Using ensemble learning and interpretable machine learning to assess flash flood susceptibility - A GIS-based Watershed Similarity Analysis
Zheng Guan[1,2], Xiaoxiang Zhang[1,2] and Jing Yao[3,4]
Abstract
This research aims to assess flash flood susceptibility, to identify watersheds with similar susceptibility risks, and to analyse the contribution of conditioning factors in Jiangxi Province, China, using four tree-based ensemble learning approaches and an interpretation tool (SHapley Additive exPlanations, SHAP). Results show that flash flood susceptibility areas in Jiangxi Province are scattered and the ensembles can efficiently perform flash flood susceptibility mapping. A comprehensive assessment of flash flood susceptibility will provide direction for flood prevention and mitigation, as well as obtaining the factors that have the greatest impact on flash floods in order to prevent them in terms of their formation mechanisms.
Presented by
Xiaoxiang Zhang
<xiaoxiang@hhu.edu.cn>
Institution
1. College of Hydrology and Water Resources, Hohai University, Nanjing, China;2. Center for Geospatial Intelligence and Watershed Science (CGIWaS), Hohai University, Nanjing, China;3. Urban Big Data Centre, University of Glasgow, Glasgow, UK;4. Centre for Sustainable Healthy and Learning Cities and Neighbourhoods, School of Social and Political Sciences, University of Glasgow, Glasgow, UK
Keywords
Flash flood susceptibility, Tree-based Ensemble Learning, Machine Learning, GIS, SHAP
Dynamic Flood Impact Assessment on Urban Road Networks A case study of Beijing, China
Yimeng Liu, Alistair Ford, Richard Dawson, Saini Yang
Abstract
Urban transport networks are vulnerable to surface water flooding, leading to impacts on economic activities, social well-being and the environment. This paper describes a flood-impact-assessment method to comprehensively assessed the economic impacts of traffic disruption from flooding in terms of time delay, fuel consumption and pollutant emission. Urban flooding in Beijing is simulated using the CADDIES-2D model, and impacts are propagated onto the transport network for assessment using the SUMO agent-based transport model. Comparing the economic damage of the baseline traffic scenario with that of three flooded scenarios, it’s demonstrated that rainfall occurring at 7 a.m. induces four times more cost than the baseline. The rain of the same intensity and duration occurring at 8 or 9 a.m. lead to a cost increase for 37.33% and 13.21% respectively.
Presented by
Yimeng Liu
Institution
School of Engineering, Newcastle University
Other Affiliations
Beijing Normal University, Beijing
Keywords
urban flooding; impact assessment; traffic simulation
Understanding Urban Traffic Flows in Response to COVID-19 pandemic with Emerging Urban Big Data in Glasgow
Yue Li, Qunshan Zhao and Mingshu Wang
Abstract
This research applies spatial Durbin model to analyse traffic flow distributions via various factors in the urban areas and traffic flow data. The results show that the overall built environment within a buffer area has more significant impact on urban traffic flow compared to the nearby location within a few meters. Areas with more young and white dwellers are associated with more traffic flows. With the influence of COVID-19, residents prefer to spend their daily life in their local neighborhood rather than having long distance travel. The initial findings from this research provide evidence of developing 20- minute city via active travel for achieving net-zero carbon target.
Presented by
Yue Li
Institution
Urban Big Data Centre, University of Glasgow
Other Affiliations
School of Geographical and Earth Sciences, University of Glasgow
Keywords
Traffic flow, Urban big data, COVID-19, Spatial Durbin model