Search site

School of Geography

Alison Heppenstall Professor Alison Heppenstall

Contact details

Room Manton 10.115
School of Geography
University of Leeds
University Road
Leeds LS2 9JT   UK


0113 343-3361

Student hours:
Please note that I do not work on Wednesday's.

Work in progress

Current Projects

I hold an ESRC-Alan Turing Fellowship entitled 'Bringing the Social City to the Smart City'.  Outline details can be found here.  A project website will shortly be set up.  Please contact me for further details.

Previous projects

 Geo-spatial restructuring of industrial trade (ESRC Funded)

a)      Create an input-output (IO) matrix between the 133 NUTS-3 UK zones and the economic activities (disaggregated by SIC 2007) that operate within them.

b)      Demonstrate the utility of the IO matrix by examining an applied question: how will changing fuel costs alter the UK’s economy spatially?

c)      Explore the implications of the trade flow and dynamic fuel cost model.

Modelling consumer behaviour: ESRC funded

The primary objectives of this research are:

  • Creation of an abstract framework for modelling individual consumer behaviour and an implementation of the framework for modelling.
  • The addition of realistic behaviours to the model for modern retail consumers using a combination of available qualitative and quantitative studies.
  • Application of the model to test-case scenarios and analysis of outputs, both to enhance our understanding of the socio-economics of the test-cases and model applicability and validation more generally.

Emerging sustainability: EPSRC funded

To develop a rigorous conceptual framework which will provide knowledge of the conditions and processes which facilitate or mitigate against the emergence of sustainability. To develop research methodologies which reflect an understanding of knowledge as an emergent property of interactions amongst humans and between humans and the systems which they observe and shape.


  • To utilise theoretical and methodological insights from complexity science in outlining a conceptual framework of sustainability as an emergent property of complex adaptive human eco-systems.
  • To utilise the principles of open-source knowledge creation to bring together researcher and practitioner knowledge from a range of disciplines and domains to enable the emergence of a coherent understanding of sustainability.
  • To undertake studies within specific domains to provide substantive examples for informing the research aims. These studies will identify the factors which mitigate against the emergence of sustainability and identify the conditions from which sustainability is emergent in specific instances.
  • To develop and refine strategies and interventions that have the potential to create the conditions necessary for sustainability to emerge, while averting catastrophic systems failure. These strategies and interventions would be tested in future research.

For further information see: Emerging Sustainability

Spatially embedded complex systems engineering (SECSE) project: EPSRC funded

A three-and-a-half year project spanning neuroscience, AI, urban geography, and complex products systems that will focus on the role of the spatial organization of networks and the spatial processes within which they are embedded in the domains of complex networks modelling, neuroscience, geo-information systems and distributed IT products such as air-traffic control..
I am currently working on a various parts of this project:

  • Development of spatial recurrence plots for visualising complex behaviour.
  • Application of agent models for modelling complexity (bifurcations, perturbations); specifically retail markets.
  • Examination of variability/Sensitivity in parameter space - neuroevolutionary methods.
  • Understanding behaviour in complex systems; identifying periodicity etc.
  • Robustness of networks (social and ecological).

Development of neuroevolutionary methodologies for modelling flood events and sediment concentrations.

This work focuses on the development of neuroevolutionary methods, specifically neural networks and genetic algorithms (JavaSANE) for improving predictions of flood events and sediment loads. Research to date has benchmarked JavaSANE against traditional neural networks, experimented with a variety of objective functions and introduced time correction functions and penalities.
See publications page for further details.
This work is in conjunction with Dr Linda See (Leeds), Dr Bob Abrahart (Nottingham) and Dr Christian Dawson (Loughborough).



08/12 Geo-spatial restructuring of industrial trade. PI: Alison Heppenstall; Co-I: Dr Gordon Mitchell. Prof Malcolm Sawyer, Dan Olner.  £179K.  ESRC Secondary Data Analysis Initiative.

04/09 Evaluating the use of Agent-Based Modelling for Synthetic Population Data.  RGS-IBG Small Research Grant. £2000.  PI: Alison Heppenstall

07/07 The GENESIS Project: GENerative E-Social Science: PI Prof M. Batty. CO-I: Alison Heppenstall, Dr Mark Birkin, Prof Paul Longley, Dr Anthony Steed, Prof MC Clarke, Prof J. Xie and Prof Sir Alan Wilson. £1.8million.

06/07: NEeis Grant. Alison Heppenstall, Dr Mike Bithnell (University of Cambridge) and Dr James Brasington (University of Aberystwyth) Awarded £8,000.

06/07: Modelling Individual Consumer Behaviour. ESRC First Time Grant. PI: Alison Heppenstall Due to start Jan 09. Awarded £200,000.

10/06: Emerging Sustainability. EPSRC. PI: Sarah Bell (UCL). Co-I: Alison Heppenstall (Leeds), Cristina Cerulli (Sheffield), Robin Durie (Exeter), Frances Griffiths (Warwick), Tamsin Higgs (Stirling), Angela Espinoa (Hull). Awarded £400,509.

10/06: Learning and Teaching Enhancement Fund Application, University of Leeds: Podcasting/Vodcasting for Environmental Scientists and Geographers. Alison Heppenstall and Lee Brown. Awarded £1,400.

Previous projects

PhD research: application of hybrid intelligent agents to modelling a dynamic, locally interacting retail market

The emergence of agent-based modelling from the field of artificial intelligence (AI) presents a new and alternative approach to geographical modelling. The vast potential offered by agent-based models in representing distributed complex systems, coupled with the increase in available computing power has resulted in agent-based models becoming an increasingly popular and powerful tool within geographical applications. These models offer distinct advantages over traditional empirical techniques through their characteristics of autonomy, flexibility and adaptability. There is an emerging recognition that the power of agent-based systems is enhanced when integrated with other AI-based and conventional approaches. The resulting hybrid models are powerful tools that combine the flexibility of the agent-based methodology with the strengths of more traditional modelling. This research examines the application of a hybrid agent-based model to the case study of the retail petrol market. Detailed analysis of the real data was first performed before the construction of an agent-based model. Model performance was evaluated against real data from the UK for a three month period in 1999. On the basis of this evaluation, the agent model was further developed to incorporate consumer behaviour by the inclusion of a spatial interaction (SI) model and a network model. Suitable parameters for these models were derived through detailed analysis of the real data, numerical experimentation and experimentation on the real data. These developments improved the performance of the model. A genetic algorithm (GA) was constructed to provide an objective approach to deriving optimal parameters. There was a close agreement in the values selected by the GA and those derived by hand. This research clearly demonstrates that agent-based modelling has the ability to improve on existing geographical models. Further investigation is needed if this potential is to be fully realised for a range if geographical problems.

MSc Research: investigating relationships in water quality using artificial intelligence and statistical techniques

An investigation was made of the potential use of artificial intelligence techniques to identify relationships within water quality. The traditional statistical techniques of K-means cluster analysis and linear regression were employed for (i) comparing how well the artificial neural network (ANN) had clustered and (ii) determining if clustering was the most suitable technique for this data. The research focuses on two pollutants, one heavy metal (zinc) and one PAH (benzo(b)fluoranthene). The ANN was found to be the most appropriate technique, with the K-means cluster analysis picking out extreme cases and the linear regression giving a poor fit to the scattered data. The relationship between environmental factors and the indices was established, as well as the observation that assessment of individual pollutants was not enough. These conclusions were used to build a predictive model. Combinations of the data set were fed into the model and the accuracy and sensitivity analysed. The results showed that a combination of two heavy metals and two PAHs gave the most accurate prediction and that the model was robust to the choice of data and within small errors in the data.