Folch & Sutherland - Using Linear Tracking Data to Estimate Backcountry Recreation Popularity

Discover how linear tracking data from Strava and AllTrails is used to estimate backcountry recreation popularity in Arizona, aiding in the identification of abandoned mines posing risks to the public.

Key takeaways
  • The project aims to estimate backcountry recreation popularity in Arizona using linear tracking data to identify abandoned mines posing risks to the public.

  • The study focuses on hiking, biking, and running activities, with plans to include hunting and off-roading in the future.

  • Data sources include social media apps (Strava and AllTrails), official GIS data, and raster sources for explanatory variables.

  • Linear data from Strava and AllTrails is converted into presence points for MaxEnt modeling.

  • Rasterization is used to address double-counting and density issues, with a recursive subdivision process to prevent segment splitting.

  • Explanatory variables are grouped into three categories: activity, ecology, and attractiveness.

  • The model uses elastic net regularization, combining lasso and ridge regression, and includes transformations for nonlinear relationships.

  • Model performance is evaluated using area under the curve (AUC) and precision-recall curve metrics.

  • Challenges include data fuzziness, computational constraints, and the need for qualitative analysis of mapped predictions.

  • The project contributes to a risk assessment that will rank abandoned mines based on their hazard level.

  • Future work involves refining methods, testing higher resolutions, and incorporating additional data sources.