How concentrated disadvantage moderates the built environment and crime relationship on street segments in Los Angeles

Criminological theories have posited that the built environment impacts where crime occurs; however, measuring is difficult. Furthermore, it uncertain whether differentially in high-disadvantage neighborhoods. This study extracts features of from Google Street View images with a machine learning semantic segmentation strategy to create measures fences, walls, buildings, and greenspace for over 66,000 street segments Los Angeles. Results indicate presence more buildings on segment was associated higher rates had particularly strong positive relationship robbery motor vehicle theft low-disadvantage Notably, fences walls exhibited different relationships crime. Walls, which do not allow visibility, were strongly negatively related crime, burglary Fences, fewer robberies larcenies, but burglaries aggravated assaults. Fences only negative violent when they located The results highlight importance accounting surrounding level disadvantage exploring micro-location