Over the past decade, sensor networks have been proven valuable to assess air quality on highly localized scales. Here we leverage innovative sensors to characterize gaseous pollutants in a complex urban environment and evaluate differences in air quality in three different Los Angeles neighborhoods where oil and gas activity is present. We deployed monitors across urban neighborhoods in South Los Angles adjacent to oil and gas facilities with varying levels of production. Using low-cost sensors built in-house, we measured methane, total non-methane hydrocarbons (TNMHCs), carbon monoxide, and carbon dioxide during three deployment campaigns over four years. The multi-sensor linear regression calibration model developed to quantify methane and TNMHCs offers up to 16% improvement in coefficient of determination and up to a 22% reduction in root mean square error for the most recent dataset as compared to previous models. The deployment results demonstrate that airborne methane concentrations are higher within a 500 m radius of three urban oil and gas facilities, as well as near a natural gas distribution pipeline, likely a result of proximity to sources. While there are numerous additional sources of TNMHCs in complex urban environments, some sites appear to be larger emitters than others. Significant methane emissions were also measured at an idle site, suggesting that fugitive emissions may still occur even if production is ceased. Episodic spikes of both compounds suggested an association with oil and gas activities, demonstrating how sensor networks can be used to elucidate community-scale sources and differences in air quality moving forward.