Climbing crags recommender system in Arco, Italy: a comparative study

Area of Science:

• Recommender Systems
• Human-Computer Interaction
• Outdoor Recreation Analytics

Background:

• Outdoor sport climbing is popular in Northern Italy, but information overload regarding climbing crags poses a challenge for tourists.
• Existing recommender systems for climbing crags often overlook crucial contextual information (season, accessibility, family-friendliness).
• Contextual factors significantly influence the feasibility and desirability of visiting outdoor activity locations.

Purpose of the Study:

• To develop and evaluate a novel context-aware recommender system for sport climbing crags.
• To address the limitations of existing systems by incorporating contextual information and user preferences.
• To improve the information sufficiency and satisfaction of climbers planning outdoor sport climbing vacations.

Main Methods:

• Collected and analyzed crag visit data from an online guidebook in Arco, Italy.
• Developed the Visit & Climb system, which learns user preferences and contextual tastes from visit logs.
• Evaluated the system through offline metrics (MAP, Recall, NDCG) and online user studies with experienced climbers.

Main Results:

• Offline evaluations showed Visit & Climb performs comparably to classical recommender models for top-N recommendations.
• Online evaluations with 40 experienced climbers indicated Visit & Climb significantly improved information sufficiency.
• The system's transparency, derived from incorporating contextual information, enhanced user satisfaction and intention to use.

Conclusions:

• Context-aware recommender systems, like Visit & Climb, are effective for outdoor activity recommendations.
• Incorporating contextual information and crag characteristics improves recommendation transparency and user satisfaction.
• The Visit & Climb system offers a valuable tool for climbers seeking personalized crag recommendations.