Horizon CDT Research Highlights

Research Highlights

Understanding User Interaction with Geospatial Features to Improve Walking Route Recommendation

  James Williams (2020 cohort)

People often use navigation systems to get assistance when looking to explore new and unfamiliar places. Existing systems focus on providing the user with the most direct and fastest route with limited attempts at providing recommendations beyond this. Recently, there has been an escalating interest in how such route recommendations can be provided to users for more leisurely activities such as walking, hiking, or running. An increase in the usage and usefulness of location-based services provides a number of application areas for such recommendations to be delivered, namely activity tracking applications such as OS Maps and Strava allow users to share routes with followers. While these location-based social networks provide a large platform for the capture and sharing of routes, limited suggestions are performed due to a lack of knowledge surrounding what features an individual interacts with whilst walking. 

Understanding user interactions with more pedestrian routes and points of interest is a major problem in providing recommendations for activities such as walking. The PhD study will therefore investigate how interactions with routes features can provide improved detail to be applied in walking recommendations. Researchers have not treated the area of more discrete interactions with geography in much detail, instead focusing on actions in the most popular areas such as city centres, where users are more likely to check-in and leave a large data footprint. These existing studies focus on using visitors as a key value of popularity, limiting the potential for knowledge to be gained of lesser-known features and how interactions with these can be used. The research proposed for this PhD will look to address this gap in knowledge, by investigating the following research question in detail:

How can understanding mobile user interactions and motion patterns support feature-based geospatial data in improving walking route recommendation?

The question above is split into three main aims defining the focus of the study, these are as follows:

  1. Understanding the User. To form an understanding of when, how, and why users interact with geospatial features while using walking routes.
  2. Understanding the Data. To establish knowledge on how scalable data can be generated from existing datasets when combined with user interactions.
  3. Design and Test. To propose, develop, and test a route recommendation system framework which uses mobile user interactions to suggest better personalised routes.

The study may provide both an academic and real-world impact, with possibilities for the knowledge gained throughout the study and proposed framework to be applied in novel recommendation systems and provide a better understanding of interactions whilst walking. This short summary highlights the status of the project at present; however, this may change in future iterations of the work.