/

• 1. Perceived versus actual unsafe areas and the link to people’s activity spaces.

  Objective: To examine whether spatial patterns of daily routes are associated with the gap between perceived and actual unsafe areas.

  Keywords: Spatial statistics, crime analysis, perception gaps, population at risk

  Acquire, develop, and apply knowledge on:

  • Spatial Groupings
  • Point Pattern Analysis
  • Logistic Regression
  • Statistical Testing
• 2. Social Media and Location Privacy: what can tweets reveal?

  Objective: To develop an algorithm that assesses inference attacks from semantic place-based signatures compared to inference attacks from spatial signatures; both derived from LBSN data. The motivation and goal of this topic is to raise awareness of the hidden risks to individual privacy by the extensive use of LBSNs.  

  Keywords: inference attacks, location privacy, semantic signatures, location-based social networks

  Acquire, develop, and apply knowledge on:

  • Python or R programming
  • Processing of LBSN data
  • Location inference algorithms
  • Geo-localisation algorithms
  • Data quality
• 3. Data protection of health geodata and its impact on spatial analysis.

  Objective: To investigate different strategies for privacy protection of health geodata and asses their effects when performing spatial and statistical analysis.

  Keywords: data protection, location privacy, health geography, spatial analysis, spatial error

  Acquire, develop, and apply knowledge on:

  • Data protection strategies
    
  • Literature on health geography
    
  • Anonymization techniques
    
  • Spatial statistical analysis
• 4. k-Voronoi Masking: a method to protect confidential discrete spatial data.

  Objective: To develop a spatial algorithm that anonymizes confidential discrete spatial data by mitigating the re-identification risk and the confidence to identify.

  Keywords: location privacy, geoprivacy, geographical masking, k-anonymity, spatial error

  Acquire, develop, and apply knowledge on:

  • Python and ArcPy package
    
  • K-anonymity algorithms
    
  • Spatial statistical analysis
• 5. Short Vs Long term spatial crime forecasting using autoregressive models.

  Objective: To evaluate the accuracy of short-term versus long-term forecasting of crime in space when applying autoregressive models.

  Keywords: predictive policing, forecasting, spatio-temporal statistics, autoregressive models

  Acquire, develop, and apply knowledge on:

  • Predictive policing
    
  • Spatial regressive models
    
  • Autoregressive models
    
  • Space-time regressive models
    
  • R Programming
• 6. Exploring Space-Time Autoregressive models for the prediction of future crime hotspots.

  Objective: To test and evaluate the ability of space-time autoregressive models to forecast crime hotspots.

  Keywords: predictive policing, forecasting, spatio-temporal statistics, autoregressive models

  Acquire, develop, and apply knowledge on:

  • Predictive policing
    
  • Spatial regressive models
    
  • Autoregressive models
    
  • Space-time regressive models
    
  • R Programming
• 7. Introducing spatial dependence in machine learning models.

  Objective: To develop a geographical modelling approach that accounts for spatial autocorrelation by introducing spatial features to traditional machine learning models.

  Keywords: geographical modelling, machine learning, predictive analytics, spatial autocorrelation, spatial dependence.

  Acquire, develop, and apply knowledge on:

  • Spatial Statistics
    
  • Supervised learning
    
  • Validation and performance evaluation
    
  • Python programming