AI and Geodemographics

Description

This research project will advance the field of geodemographic classification by developing novel methodological approaches to integrate and analyse diverse urban data sources. The increasing availability of structured socioeconomic data alongside rich unstructured visual information presents unprecedented opportunities to enhance our understanding of residential and commercial areas through computational methods.

The research will develop an innovative multimodal framework that synthesizes traditional demographic and economic indicators with street-level imagery and advanced AI technologies. Through the application of deep learning architectures for feature extraction from visual data, combined with large language models for generating interpretable descriptions, this work aims to create more nuanced and accessible geodemographic classifications.

The successful candidate will develop and validate sophisticated methodological approaches for data fusion, training advanced AI models to extract, integrate, and represent urban characteristics across multiple modalities. A key research objective is the generation of interpretable visualisations and narrative descriptions of classifications, ensuring that outputs are both methodologically robust and accessible to diverse stakeholders including policymakers and urban planners. Through rigorous empirical validation and detailed case studies, the candidate will demonstrate the framework's capacity to generate actionable insights into urban structure and function.

This project positions the researcher at the intersection of artificial intelligence and urban analytics, combining methodological innovation with significant potential for practical impact. The successful candidate will contribute to advancing the theoretical foundations of AI-driven geodemographics while developing tools and approaches that can inform evidence-based urban policy and planning decisions.

Training and Collaboration

This project sits within the Geographic Data Service (www.geods.ac.uk), which is a national data service that is funded as part of UKRI Smart Data Research UK. The mission of our service is to maximise the value of diverse and nationally significant Smart Data through geographic integration, enrichment and validation. We support and empower our data service users to generate location-aware insight and impact for projects in the public good. As such, this PhD is part of a large team based at the University of Liverpool that is working on related data science projects or development of supporting software service infrastructure. You will also be part of the Geographic Data Science lab, which hosts a large number of other research projects and PhD students and will benefit from both cohort support and a wide range of lab activities such as seminars, workshops, peer review and software demonstration. Beyond these indirect training and collaboration opportunities, the support package associated with this PhD also includes a support budget of £5,000 over the duration of study.

Minimum and Desired Qualifications

Minimum Qualifications

1. Educational Background:
• A strong undergraduate degree (high 2:1 / 1^st) in a relevant field such as Geography, Data Science, Computer Science, Urban Studies, or a related quantitative discipline.
2. Technical Skills:
• Proficiency in programming languages such as Python or R, particularly for data analysis and machine learning applications.
• Familiarity with machine learning frameworks (e.g., TensorFlow, PyTorch, or Scikit-learn).
3. Analytical Skills:
• Basic knowledge of geospatial data analysis, including working with GIS platforms or geospatial libraries (e.g., GeoPandas).
4. Communication Skills:
• Ability to clearly present research ideas and findings both in writing and verbally.
5. Research Experience:
• Some experience with data handling and statistical methods, preferably through coursework or projects.

Desired Qualifications

1. Advanced Educational Background:
• A master’s degree in a relevant field, such as Computer Science, Geographic Data Science, Artificial Intelligence, Data Science or Urban Analytics.
2. Deep Learning Expertise:
• Experience with advanced machine learning techniques, including autoencoders, convolutional neural networks (CNNs), or multimodal models.
3. Geodemographic Knowledge:
• Familiarity with geodemographic methods and concepts, particularly using structured data like census or survey datasets.
4. Generative AI Experience:
• Experience working with large language models (e.g., GPT, BERT) for natural language generation or retrieval-augmented generation tasks; and frameworks (e.g. HuggingFace/SpaCy).
5. Image Processing Skills:
• Knowledge of image analysis techniques, including preprocessing and feature extraction using computer vision frameworks.
6. Geospatial Data Expertise:
• Advanced skills in handling geospatial datasets, spatial statistics and geospatial database management (e.g., PostGIS).
7. Interdisciplinary Insight:
• Understanding of urban planning or sociology to contextualise technical outputs within broader academic frameworks.
8. Project Management and Collaboration:
• Demonstrated ability to manage complex projects and collaborate in interdisciplinary teams.
9. Publication Experience:
• Previous experience in publishing research findings in academic journals or presenting at conferences.

Deadline Changes

Please note that applications for this position may close earlier than the stated deadline if a suitable candidate is identified. We encourage interested applicants to submit their materials as early as possible to ensure consideration. Any updates to the application deadline will be communicated promptly, but applicants are advised to prioritise early submission to avoid missing this opportunity.