Describing Explored Places through OpenStreetMap Data

Research Background and Issues

• Identified Problems or Challenges: Existing mobile navigation applications primarily focus on providing the shortest and most efficient routes, which restricts user exploration behavior. This results in large areas of urban spaces being underexplored, thereby impacting users' spatial learning.
• Significance: Exploratory behavior is crucial for human spatial cognition, learning urban layouts, and building cognitive maps. Relying solely on the shortest routes may weaken users' ability to acquire spatial knowledge.
• Research Motivation and Related Work: Although some studies have proposed alternative navigation routes based on aesthetics or safety, these route designs often fail to fully incorporate actual user exploration data. This study aims to define the characteristics of attractive locations by analyzing real data from users' free exploration, thereby improving navigation algorithms.

Solution

• Proposed Method or Solution:
  • Collect and analyze user trajectory data from a gamified mobile application (MapUncover).
  • Use OpenStreetMap (OSM) data tags and a tile-based mapping method to describe the characteristics of locations explored by users.
• Innovations:
  • Utilize data from real user exploration processes rather than developer-assumed efficiency or aesthetic optimization routes.
  • Propose a method for analyzing OSM tags to provide an objective basis for studying exploration behavior.
  • Compare the characteristics of freely explored locations with those accessed via traditional shortest-path algorithms.
• Implementation Steps and Key Techniques:
  1. Data Collection: The [MapUncover] application incentivized users to engage in spatial exploration through gamification, collecting data from 39 participants who unlocked 12,971 tiles over 455 days, covering an area of 106.5 square kilometers.
  2. Data Processing: Convert user trajectory data into a tile-based representation and integrate it with OSM's Nominatim geocoding API to extract the address, category, and type of each tile.
  3. Category and Type Distribution Analysis: Summarize the main characteristics of frequently explored locations and compare exploration paths with simulated shortest-path results to evaluate differences.
  4. Frequency Analysis and Weighted Distribution: Adjust category weights based on tile visit frequency to reveal tiles that are more attractive to users.

Research Findings

• Specific Findings:
  • Identified location types frequently visited during free exploration, such as office spaces, educational institutions (universities, schools), retail, commercial buildings, and tourist areas.
  • Locations explored by users differ significantly from those on shortest paths, with free exploration including more educational, cultural, and recreational areas.
  • Experimental results show that retail and educational institutions are more attractive to explorers compared to green spaces (e.g., parks), which conflicts with current navigation applications' emphasis on clean and green routes.
• Comparison with Existing Solutions and Advantages:
  • Traditional shortest-path-based navigation decisions are relatively one-dimensional, whereas this study reveals that locations outside the shortest path (e.g., educational and historical buildings) may be more appealing.
  • Most existing navigation systems fail to adequately address users' need to explore urban diversity. This study provides theoretical and data-driven support to optimize such issues.
• Experimental or Evaluation Results:
  • In data from the German cities of Hamburg and Bremen, exploration paths covered 10.93% of Hamburg's dense urban area and 31.01% of Bremen's. Exploration locations tended to avoid areas concentrated along short paths.
  • Compared to simulated shortest paths, educational institutions and cultural landmarks (e.g., universities and museums) were significantly more emphasized in free exploration.
• Limitations and Future Directions:
  • Geographic and cultural bias: The study's data is primarily sourced from central Germany, and its applicability to other countries or cultures remains to be validated.
  • Free exploration behavior may be influenced by the gamification mechanisms of the application, which might not fully reflect users' natural behavior patterns.
  • Data coverage is mainly focused on urban environments; future research should validate findings in rural or smaller city settings.
  • The impact of gender or safety factors on pedestrian navigation has not been thoroughly explored, necessitating further research to uncover more personalized routing preferences.