DescriptionExtended exposure to air pollution is a health hazard. Avoiding polluted areas to the extent possible can minimize long-term pollution exposure. To this end, users could reduce exposure to pollution by accessing an accurate air quality information stream. The air quality inventory currently available to end-users is coarse-grained (measure- ments spatially and temporally few and far between) and cannot be relied upon to make informed decisions regarding pollution exposure. Considering the hazards associated with prolonged exposure to air pollution, it is desirable to provide fine-grained (spatially and temporally dense measurements) pollution information to a variety of end-users, allowing them to make better-informed decisions using a new, more accurate information stream. This dissertation aims to fill this need by proposing new recipes to achieve fine-grained information streams with pollution inventory. In the first part of the dissertation, we present two mobile sensing platforms for fine-grained real-time pollution measurements - a portable sensing platform deployable on public transportation infrastructure and a personal sensing device that can create a social pollution sensing network. We conclude that mobile sensing platforms deployed on public transportation infrastructure can help collect fine-grained pollution measurements. We also found that a personal sensing device conveniently mounted inside a vehicle in front of the vent can measure CO levels that correlate well with outdoor values. Based on this, we propose a new neural network model, ”InsideOut,” to use the CO measurements collected by the personal sensing device inside a car to infer the CO measurements outside the vehicle, thus contributing to the fine-grained pollution inventory outdoors. In the second part of the dissertation, we propose ”X-PoSuRe” - a neural network-based regression model for pollution super-resolution trained to infer fine-grained pollution information from coarse-grained pollution measurements akin to image super-resolution, where a neural network model creates high-resolution images from low- resolution images. The X-PoSuRe model uses Nitrogen Dioxide (NO2) as the pollutant of choice and uses other covariates like meteorological data, traffic data, construction activity, accident information, large-scale fire incidents, and building footprint infor- mation for pollution super-resolution. The model is made using an ensemble of neural network models and trained using a gradient boosting technique. The results show that this model has high accuracy in inferring fine-grained NO2 concentrations from the co-variates. The proposed X-PoSuRe model provides a promising new and novel method for pollution super-resolution from existing low-resolution data sources without the need for deploying expensive measurement equipment over a large area. As a next step, the X-PoSuRe model is extended to infer fine-grained measurements for Nitrogen Dioxide (NO2), Black Carbon (BC), Ozone (O3), and Carbon Dioxide (CO2) using the same covariates used initially. Even though a single inference model for multiple pollutants was not feasible, multiple inference models could be efficiently derived through transfer learning using the neural network architecture and the gradient boosting learning technique. In the third and final step, we evaluate a healthy route recommendation schema that uses these fine-grained pollution measurements to recommend healthy route options that avoid polluted road segments. These healthy routes were assessed on a neighborhood scale using measurements from mobile sensor assemblies mounted on vehicles. Experiments demonstrate that a significant reduction in pollution exposure can be achieved by choosing a healthy route instead of the shortest or quickest route.