Recent　studies　empirically　support　the　role　of　the　built　environment　in　inducing　or　hindering　violent　crime.　Particularly,　studies　of　the　broken　window　theory　have　provided　evidence　that　physical　disorder　is　an　environmental　correlate　of　crime.　This　includes　broken　windows,　vacant/abandoned　housings,　abandoned　cars　on　street,　graffiti,　and　decayed　street　lighting,　among　other　things.　Current　studies　are　limited　by　the　difficulty　involved　in　collecting　fine-scale　quantitative　environmental　data.　The　conventional　environmental　audit　approach,　which　aims　to　assess　environmental　features,　is　costly,　time-consuming,　and　burdensome.　In　this　study,　we　use　Google　Street　View　to　study　the　relationship　between　violent　crime　and　physical　features　of　urban　residential　environment.　More　concretely,　a　Poisson　regression　model　with　spatial　filtering　is　used　to　identify　socio-economic　correlates　of　violent　crime.　Parting　from　the　hypothesis　that　omission　of　built　environmental　factors　results　in　systematic　residual　pattern,　we　proceed　to　analyze　the　spatial　filter　to　select　sites　for　virtual　environmental　audits.　A　series　of　physical　environmental　factors　are　identified　using　contingency　table　analysis.　The　results　provide　both　theoretical　and　practical　implications　for　several　theories　of　crime　and　crime　prevention　efforts.　(C)　2017　Elsevier　Ltd.　All　rights　reserved.