Congratulations to our winners this year. The SLA Student Awards 2023 reached new goals this year with 9 entries from the following universities: Leeds, University College London and Birkbeck University of London.
We awarded two prizes in each category; Masters and Undergraduates.
Masters Award winners 2023
Our two Masters winners were:
- Elliot Moses – Birkbeck, University of London
- Liyuan Dong – University College London
Undergraduate Award Winners 2023
Taking home the prizes in the Undergraduates category were:
- Jan Magnuszewski – University of Leeds
- Emma Andre – University of Leeds
Meet our winners
Masters Award 2023
1st Place – Eliott Moses, Birkbeck, University London
Dissertation Title: ‘Exploring areas of upgrade based on the 15-minute city concept – A case study of London Using Spatio-temporal Analysis and Machine Learning to Predict Transport for London Bike-Sharing Habits in the Post COVID-19 era.’
Dissertation Synopsis – Bike-sharing systems offer a convenient option to efficient and sustainable urban transportation, alleviating congestion. The growing demand necessitates continued understanding of how the systems are used. This study examines the shifts in Transport for London bike-sharing habits during the post COVID-19 era through exploratory and statistical analyses, focusing specifically on bike journeys undertaken in 2019 and 2022, totalling nearly 22 million rides. Employing network analysis, significant docking stations were identified, often positioned near locales of work, transport, and recreation. The overall usage saw a rise in 2022. Consistent with previous findings, weather and summer months exert a strong influence on bike share usage patterns. Temporal signals have reverted to pre-pandemic norms. Yet, a noticeable decrease in commute-hour rides alludes to the ongoing evolution of working habits. Moreover, there was a rise in weekend rides, markedly at docking stations where leisure usage dominates, such as Hyde Park Corner.
For current and future systems to seamlessly integrate into the urban transport network, achieving precise forecasts for travel demand at a detailed spatio-temporal level is paramount. This research also examines the influence of temporal and meteorological factors on bike-sharing demand. Leveraging the Long Short-Term Memory neural network model, hourly demand was predicted. The model demonstrated effective prediction capabilities, particularly at docking stations where specific user types dominate, such as commuters. These insights bolster sustainable transport strategies and forecasting for bike-sharing systems.
Research Questions
- How were TfL bikes used before COVID-19 (2019)? Briefly confirm findings found in previous research, analysing data from 2019 and 2020.
- How are TfL bikes being used in the post COVID-19 era (2022)? Establish changes analysing data from 2022, if any, to cycling habits.
- To what extent can Machine Learning be employed to predict TfL’s bike-sharing habits in the future? Using data from 2022, make predictions on usage.
2nd place – Liyuan Dong – University College London
Dissertation Title: ‘A comparative study of geographically weighted regression methods for house price modelling based on the London study case‘
Dissertation Synopsis – The main focus of this project is on the comparison of traditional geographically weighted regression and, emerging methods combining geographically weighted regression and machine learning, exploring the differences and advantages and disadvantages between these methods based on London house prices as a case study. As global house prices continue to climb, the property market is generating more and more concern. Consequently, house price prediction has become a heated topic of discussion. As more and more methods are proposed and developed, the results of house price prediction are gradually optimised. Simultaneously, when studying the topic of house prices, it is unavoidable to consider the geographical situation of the house. Therefore, the method of geographically weighted regression (GWR) is applied to predict house prices. In addition to linear models for prediction, some non-linear methods, machine learning methods, have been applied to predict the house prices, such as random forest (RF), and artificial neural network (ANN). Recently, some scholars have proposed and implemented a combination of geographically weighted ideas and machine learning derive a geographically weighted non-linear algorithm. Thus, Hagenauer (2022) presented the geographically weighted artificial neural network (GWANN) and Georganos et al., (2021) presented the geographically weighted random forest (GWRF) to address predictions with spatial effects. Moreover, GWANN has been applied to predict house prices in Austria but GWRF has only been applied to predict population modelling but not house prices. Therefore, as the emerging methods GWANN, GWRF have been developed only recently, his study will be the first study to compare the performance of GWR, GWANN, and GWRF using London house prices as a case, to conclude which model is optimal as well as their strengths and weaknesses, and to compare the differences in the predictive effectiveness of geographically weighted and non-geographically weighted methods.
Undergraduate Award 2023
1st place – Jan Magnuszewski, University of Leeds
Dissertation Title: ‘The Kiwi Urban Liveability Index: Analysing urban liveability in Auckland, New Zealand with a spatial composite indicator’
Dissertation synopsis – Enhancing urban liveability in the cities of Aotearoa New Zealand represents a shared goal for both local and central governments, seeking to make urban living more enjoyable, equitable and sustainable.
The academic discourse around the topic, largely focused on Auckland, has discussed the concept of liveability in the context of intensification policies (Beattie and Haarhoff, 2018), and through interviews, has identified the aspects that contribute to Aucklander’s perceived liveability (Haarhoff et al., 2016).
This research aims to complement this discussion quantitatively, by proposing the Kiwi Urban Liveability Index (KULI) – a spatial liveability indicator for Auckland, the need for which, has been recognised by Auckland City Council (Meares et al. 2015). The index is made up of 37 indicators, spanning the key dimensions of liveability: social infrastructure, green space, transportation, walkability, safety and culture.
The indicators were created using GIS methods including network distance calculation, and represent the aspects which make cities liveable, with several New Zealand-specific additions. Through the evaluation and geospatial analysis of the KULI, this dissertation also studies its applicability, and draws insights about the patterns of liveability in Auckland, in relation to the underlying characteristics of the population. Among the key findings are the positive correlation between liveability and use of public and active transport, as well as the finding that Pacific and Māori communities are disproportionately disadvantaged in access to liveable spaces. The findings illustrate the applicability of the KULI, demonstrating its potential for use in urban planning and research.
- The most liveable locations in Auckland, as identified with the KULI were found to broadly correspond to the main development areas outlined in the Auckland Plan 2050, which shows the agreement between the theory behind the KULI and the city’s official planning agenda.
- The multivariate regression analysis confirms the conventional notion that higher liveability is associated with higher usage of public and active modes of transport. The analysis also revealed neutral relationships between liveability and level of income and deprivation.
- By utilising geographically weighted regression models, the study provides insights into the spatial variations of these relationships. For instance, in locations with relatively low Māori population, an increase in this percentage is positively associated with liveability.
- The analysis also highlights the inequalities in living standards, between different ethnicities in Auckland, by showing that the Pasifika and Māori communities tend to have the lowest percentage of their population inhabiting the most liveable locations, compared to the Pākeha and Asian populations.
2nd place – Emma Andre, University of Leeds
Dissertation Title: ‘Siting Social Supermarkets in Yorkshire and The Humber: Can Access be Improved for Food Insecure Communities?‘
Dissertation Synopsis – The cost-of-living crisis in the UK has resulted in increased levels of food insecurity. Consequently, the supply of food aid needs to be re-examined to ensure there is enough provision for food insecure communities. Food banks have been criticised as a short-term response to food insecurity, and do not address the root causes of the problem. Social Supermarkets are an alternative form of food aid which provide social and economic support to users, in addition to food aid provision. There has been a lack of research regarding where Social Supermarkets are located and how accessible they are. Therefore, the primary aim of this research is to identify the existing coverage of Social Supermarkets to the most food insecure groups in Yorkshire and The Humber, and to determine how access coverage can be expanded optimally. This is achieved by locating communities who are at high-risk to food insecurity. A composite indicator measures demographic vulnerability to food insecurity and a Hansen Index will determine areas with a lack of grocery opportunities.
This dissertation also assesses the existing supply of social supermarkets and uses isochrone catchment analysis to determine the extent of accessibility for food insecure areas. A location-allocation model optimises the placement of 50 additional Social Supermarkets. Finally, levels of accessibility for high-risk communities are determined and compared to the model output via isochrone catchment analysis. Overall, this dissertation contributes to the understanding of food insecurity and access to social supermarkets in Yorkshire and The Humber. It provides actionable insights for improving the reach of social supermarkets and highlights the importance of addressing food insecurity disparities in both urban and rural settings.
The geospatial analysis techniques applied offer valuable tools for policymakers, practitioners, and researchers working to alleviate food insecurity challenges.
Congratulations to all our winners, we had such a great response of very good quality dissertations this year.