Low-cost particulate sensors can allow for establishing a dense monitoring network to increase the spatial resolution of air quality information, which is particularly of interest in urban areas. However, these sensors are often affected by environmental factors such as temperature and humidity, the effects of which must be accounted for so that the accuracy of these sensors in field conditions can be quantified. In this paper, we conduct long-term tests of two types of low-cost particulate sensors: Met-One NPM and PurpleAir units. We assess the self-consistency of larger groups (12 to 25) of sensors, develop empirical equations for correcting the measurements of these sensors to better match those of regulatory-grade instruments, and assess the long-term performance of these sensors during deployments lasting over a year. These assessments are used to assess sensor performance in two different use cases: improving community awareness of air quality with a focus on short-term qualitative indications and providing accurate long-term quantitative information for health impact studies. We find that, for the short-term case, using either quadratic or piecewise-linear correction equations, either sensor can be used to provide reasonably accurate concentration information for PM2.5 (mean absolute error on the order of 4 µg/m3) in near-real time. For the long-term case, by applying in-field noise-adjustment, bias can be reduced below 1 µg/m3. These results indicate the suitability of these sensors for supplementing regulatory-grade instruments in sparsely monitored regions, as well as for conducting hotspot mapping to better understand the variability of air quality in urban areas.