RGS-IBG Annual International Conference 2017

RGS-IBG Logo

358 Workshop: Spatial Urban Analytics and Crowdsourced Geographic Information for Smarter Cities (1)
Affiliation Geographical Information Science Research Group
Convenor(s) João Porto de Albuquerque (University of Warwick, UK)
René Westerholt (Heidelberg University, Germany)
Chair(s) João Porto de Albuquerque (University of Warwick, UK)
Timetable Friday 01 September 2017, Session 3 (14:40 - 16:20)
Session abstract The recent emergence and availability of ever more data reflecting everyday human behavior within urban areas opens up opportunities for geographers and strengthens the geospatial viewpoint in the interdisciplinary field of urban science. This session focuses on methods and applications based on crowdsoucing, social media data, collaborative maps (e.g. OpenStreetMap) and mobile crowd sensing/citizen science approaches, with a particular emphasis on explicitly geospatial concepts and methods to the interdisciplinary field of urban analytics. The talks span across a broad range of topics dealing with conceptual innovations, the extension of existing spatial data analysis techniques and the development of new methods that explicitly consider spatial issues (in contrast with more general, non-geographic computational methods). Aside of concepts and methods, some talks consider scenarios related to smart cities, human mobility, urban planning and others.
Linked Sessions Workshop: Spatial Urban Analytics and Crowdsourced Geographic Information for Smarter Cities (2)
Contact the conference organisers to request a change to session or paper details: AC2017@rgs.org
Welcome note
João Porto de Albuquerque (University of Warwick, UK)
René Westerholt (Heidelberg University, Germany)
Apart from welcoming speakers and audience, we provide a brief overview of all submitted contributions. In addition, in order to give the big picture, we highlight how these are related to each other and how they fit into the urban science landscape.
A platform for measuring urban functionality from social media data
Chen Zhong (King's College London, UK)
Michele Ferretti (King's College London, UK)
We present a pilot project here, which is to is to extract information about urban functions from real-time streaming of geo-referenced tweets by advanced data mining methods. The extracted information is considered as citizens’ feedback to urban development, telling us how they think, redefine and use the urban space in reality. We use these feedbacks to evaluate the original land use plans and transport policy, and see if the expected impacts (e.g. improving life efficiency, better access to shops, more compact urban space) have been achieved. Moreover, such feedbacks can be updated frequently as streaming data are generated automatically and constantly. As such, the extracted information is valuable references for stakeholders who make decisions of urban plans to adjust the pathway of urban transformation and/or regeneration accordingly.

In a broader sense, this research is in line with the new concepts about urban futures, namely ‘Smart City’, ‘Urban Big Data’ and ‘Citizen Science’. The idea behind this project is to facilitate the use of automatically generated human mobility data for understanding the interactions between people and their built environment, and eventually, to power the data-driven approach in social and urban studies. Taking Twitter data as an example, here we demonstrate that automatic data, if can be interpreted in a certain context, could give us much more frequently-updated information with finer granularity in better coverage and with less questionnaire bias, comparing to conventional surveyed approach. In particular, this project is

1. To develop a framework for any other plug-in type of data analysis and visualization applications built an information infrastructure, which completes a pipeline from data collection, storage, analysis and geo-visualization. It is to test the feasibility of our solution for converting streaming data into information in real-time, which differentiates our system from the kind of already populated pure data
mapping tools.
2. To develop fast data mining methods, which will be embedded in the data processing pipeline, and are the essential advances of our solution.
3. To conduct a case study, which is to extract information of urban functions from real-time streaming geo-referenced tweets.
The role of mobility in exploring spatial aspects of liveability using big data
Anna Kovács-Győri (University of Salzburg, Austria)
Liveability is a popular concept regarding city quality. Creating liveable places is what many urban planners and decision makers wish to achieve. Liveability measurements can foster this because they enable the exploration of interrelationships and underlying phenomena, thereby improving the efficiency and impact of implemented actions. However, two questions emerge: what is meant by using the term ‘liveability’? And how can we measure it?

The first question derives from the lack of consensus on a general definition, while every liveability-related discipline has their own practice. As a consequence, several similar, overlapping concepts are used, sometimes incorrectly and inconsistently, such as quality of life, well-being, sustainability, and urban environmental quality. Using them as synonyms impedes the proper distinction. However, the categorization of these concepts is essential for consequent liveability measurements and is more important than the precise liveability definition itself.

Applying the categorization of the aforementioned concepts, I formed a theoretical background as a starting point for further practical analyses. According to the practitioners and relevant literature, transportation (mobility) can be highlighted due to its direct and indirect role in liveability analyses. Aspects of it, such as accessibility and multimodality are inevitable factors especially in terms of pedestrians and cyclists, but according to Gehl (2010) even the necessity of a trip can be informative on city quality and thereby aids liveability assessments.

Regarding the second question, the capabilities of GIS can be relevant for liveability measurements. Quantitative (and quantifiable) spatial data can be used for each selected liveability factor as input for methods such as weighted overlay analysis. However, the level of individual subjectivity is significant, and it makes the usage of universal weights almost impossible, or at least less useful. How can we find weights for all these factors, especially on an individual level? Preferences can change even for the same person from time to time depending on ones life situation. Therefore, I tried to find an alternative which shows these preferences and can function as an indicator to decrease the number of necessary factors. Thus, I intend to test whether the relative liveability of a given place is definable using mobility analysis and whether findings based on mobility correlates with overall liveability.

For this analysis, firstly the individual trajectories should be extracted from the data, which is followed by the determination of the transportation mode. After these, based on the location frequencies, average travel time, and other possible regularities in patterns (implementing additional data sets) the trips should be located on a necessary-optional scale for each person. Finally, the evaluation on spatial level is accomplished using the statement of Gehl (2010) as a hypothesis, namely, for optional (outdoor) activities a good urban quality is a pre-requisite. For validating the results questionnaires or expert interviews are required. Once the indicator-nature of mobility is proven, I plan to use regression analysis (or equivalent method), to reveal the relevance of other liveability factors besides mobility. The general aim is to create a transferable method using spatial analysis for liveability measurements.

GEHL, J. (2010). Cities for people. Washington, DC: Island Press.
Locating the social: Embedding spatial urban analytics into the operation of critical urban infrastructure
Philipp Ulbrich (University of Warwick, UK)
Jon Coaffee (University of Warwick, UK)
In recent years, resilience ideas have extended the performance of risk management and sought to advance ways of coping and thriving in an uncertain and volatile future premised upon advancing an all-encompassing, integrated approach to engage with future uncertainty. Significantly implicated in this endeavour has been critical urban infrastructure as a result of the frequency and severity of recent crises that have channelled attention to vulnerable physical assets, the removal or suspension of which from normal service, would significantly affect public safety, security, economic activity, social functioning or environmental quality.

Until now, governance frameworks and indicator sets underpinning infrastructural operations have largely adopted a techno-scientific view, leading to engineered, “robustyet-fragile” approaches that often fail to account for the organisational, social and political factors that shape local adaptive capacity. However, more adaptive and resilient forms of critical urban infrastructure resilience call for new socio-technical approaches to better manage infrastructure in the light of known and unknown risks.

In this paper we seek to present a new methodological approach for assessing the resilience of critical urban infrastructure and obtain a better understanding of the validity and applicability of new resilience indicators across different socio-political contexts and infrastructural challenges. For this purpose we will use indicators from existing resilience assessment frameworks and apply citizen science techniques for improving data gathering and validation. Of particular interest is the formation of new associations and networks in communities that have faced disruptive or slow-burn stresses through the collation of textbased VGI in the form of citizen-generated geographic and spatial content, relating to the disruption and social processes in the affected area. Specifically, we will advance an approach that applies geo-located social network analysis complemented with participatory research methods to illuminate the networks that could emerge and assess their potential for advancing adaptive infrastructural resilience. Additionally we will focus on the macroorganisational aspects of critical infrastructure governance and its path dependency. Based on the insights from the network analysis and organisational resilience, participative methods will be used to assess to what extent these networked practices can potentially enhance institutional diversity (and hence resilience) - for example by providing a voice to previously less prominent actors.

Within this context, and drawing on the results and exemplar case studies from international projects focused on operationalising infrastructure resilience, this paper provides a critical assessment of how resilience and spatial urban analytics - so called smart resilience - shapes the ways urban infrastructure providers deal with complex risk and the tensions elicited in the paradigm shift from traditional risk management towards adaptive and more inclusive resilience approaches. Also highlighted are the implications for organisational governance in seeking holistic, integrated and more inclusive ways of assessing risk across multiple systems, networks and scales.
How Twitter and Instagram can locate ‘food deserts’ and predict cancer rate: Studying dietary choices and chronic diseases in food deprivation areas in London
Elisabeth Titis (University of Warwick, UK)
Considering high costs of health problems outcomes in terms of both healthcare and lost productivity, regions of the country likely to be or become ‘food deserts’, as well as recognising their nutritional and dietary challenges, have become increasingly im-portant to public health. Due to traditional efforts in this regard impose various limitations, especially in terms of methods (Macintyre et al., 2007), governments may collect and rely on incomplete data, with some communities being overlooked (Cendejas et al., n.d.); for more accurate empirical observations, social media has been on a rise, with ingestion content and language usage reflecting individual’s and population’s patterns of consumption (De Choudhury, Counts, & Horvitz, 2013).

Motivated by knowledge gaps in the evidence base (Wrigley, 2002), this research investigates issues of ‘food deserts’ and food poverty in the context of social exclusion and nutrition-related ill-health. The aim is to examine how social media platforms can provide empirical quantitative evidence for understanding the relation between dietary patterns and chronic disease with a particular focus on areas characterised by food deprivation, being commonly referred to as ‘food deserts’. Following hypothesis that a population in ‘food deserts’ is more likely than population in other neighbourhoods to develop chronic disease as a result of poorer food choices, this research aims a) to investigate to what extent this would be the truth, and b) to evaluate chronic disease rate, as predicted by an machine learning (ML) model. In particular, by studying dietary choices and challenges of social media users, this research seeks to answer whether, how, and to what extent food mentions on Twitter and Instagram could a) identify whether a neighbourhood is or has a potential to become a ‘food desert’, and b) capture county-wide health signals with regard to cancer and other chronic disease occurrences (CVD, diabetes type 2, obesity). The ultimate aim is to assess the adoption of Twitter and Facebook in terms of content and language usage as a way to gather better empirical evidence on the health inequalities agendas.

The focus is on mapping food poverty and ‘food deserts’ in Greater London being de-fined within the context of low-income and poor accessibility in terms of retail geography. The novelty is the combination of conventional datasets with new forms of da-ta harvested from social media platforms. Data types include a) health data, b) socio-economic deprivation data, c) geo-coded retail food sources, and d) Instagram and Twitter food-related posts with tags. A Geographical Information System (GIS) using QGIS tolls will be adopted to implement three types of spatial analysis, namely coverage, density and proximity, in order to evaluate the geographical access to healthy and affordable food of the populations in Greater London; in addition, spatial identification of ‘food deserts’ in Greater London will help to visualise areas with the highest demand for improvement in healthy food access and may be applicable to other areas with similar characteristics for strategic health interventions’ and retail planning.