Friday, April 4, 2025

Development of a multimodal semantic Knowledge Graph for GeoQA

 Suggested by: Franz Welscher, Johannes Scholz 

 

Keywords: Semantic Knowledge Graphs, GeoQA, Geospatial Data Integration, Raster and Vector Data, Spatial Querying 

 

Objective: Develop an approach for integrating raster and vector geospatial data into a semantic knowledge graph, enabling structured and efficient querying through a GeoQA (Geospatial Question Answering) engine. The research will explore data modeling, indexing, and retrieval strategies to enhance geospatial decision-making in disaster and conflict management scenarios. 

 

Short Description: This thesis investigates the integration of heterogeneous geospatial data—raster (e.g., satellite imagery, elevation models) and vector (e.g., administrative boundaries, infrastructure networks)—into a semantic knowledge graph to support advanced geospatial querying via a GeoQA engine. The study will develop and test a framework that enhances spatial reasoning, enabling users to retrieve and analyze geospatial information more effectively. A potential use case includes the PeaceEye initiative, focusing on disaster and conflict management applications. 


Start: 

As soon as possible 

 

Prerequisites/qualification: 

  • Knowledge in GIS and spatial data processing 

  • Interest/Knowledge in Knowledge Graphs 

Development of an ABM – GeoAI interaction “application” 

 Suggested by: Johannes Scholz 


Keywords: Agent-Based Modeling (ABM), GeoAI, Interactive Tools, Decision Support Systems, GeoJSON, Spatial Analysis, AI-based Recommendations, Interactive Application 

 

Objective: Develop an interactive tool that integrates Agent-Based Models (ABM) with GeoAI, enabling the system to read agent context (via post messages or GeoJSON objects) and provide AI-based recommendations for agents' next steps or decisions in real time. 

 

Short Description: This thesis will focus on creating a basic interactive application that combines Agent-Based Modeling (ABM) with GeoAI techniques. The tool will receive agent data (such as context, location, and behavior) through post messages or GeoJSON and use GeoAI algorithms to deliver recommendations for the next action or decision for the agents. The system will aim to enhance the decision-making processes in dynamic, spatially aware environments, making it useful for urban planning, disaster response, or resource management scenarios. 

 

Start: 

As soon as possible 

 

Prerequisites/qualification: 

  • Knowledge in GIS and spatial data processing 

  • Interest/Knowledge in GeoAI 

  • Interest/Knowledge in Agent-based Modelling 

Disaggregation of energy demand profiles with GeoAI methods based on sparse Smart Meter data, Socio-demographic data and mobile phone data 

 Suggested by: Johannes Scholz 


Keywords: GeoAI, Energy Demand Disaggregation, Smart Meter Data, Socio-Demographic Data, Mobile Phone Data, Machine Learning, Spatial Data Analysis, Electricity Consumption Modeling 

 

Objective: Develop a methodology for disaggregating electricity demand profiles from local supply transformers to individual households using GeoAI methods. The approach will leverage sparse Smart Meter data, socio-demographic information, and mobile phone data to enhance estimation accuracy. 

 

Short Description: This thesis focuses on using GeoAI techniques to refine the disaggregation of energy demand profiles at a granular level. By integrating Smart Meter readings with socio-demographic and mobile phone data, the research aims to model electricity consumption patterns with high spatial accuracy. The study will develop and validate machine learning-based methodologies to infer household-level demand from aggregated transformer-level data, enabling better grid management and energy planning. 

 

Start: 

As soon as possible 

 

Prerequisites/qualification: 

  • Knowledge in GIS and spatial data processing 

  • Interest/Knowledge in GeoAI 

Uncertainty evaluation in GeoAI models with sparse training data and Causal Graphs 

 Suggested by: Johannes Scholz 


Keywords: GeoAI, Uncertainty Quantification, Sparse Data, Causal Graphs, Bayesian Inference, Machine Learning, Spatio-Temporal Analysis, Environmental Modeling 

 

Objective: Evaluate the uncertainty of GeoAI models when trained on sparse datasets and analyze how uncertainty behaves with varying levels of training data quantity and quality. Additionally, assess the influence of causal graphs in improving model robustness and interpretability. 

 

Short Description: This thesis investigates uncertainty quantification in GeoAI models that rely on sparse training data, such as wildfire occurrence datasets. The research will explore how different amounts and qualities of data impact uncertainty levels and prediction confidence. Furthermore, it will examine the role of causal graphs in enhancing model understanding and reducing uncertainty. By applying statistical and machine learning techniques, the study aims to improve the reliability of GeoAI models in real-world applications. 

 

Start: 

As soon as possible 

 

Prerequisites/qualification: 

  • Knowledge in GIS and spatial data processing 

  • Interest/Knowledge in GeoAI 

  • Interest/Knowledge in Causality