Microseismic monitoring is extremely valuable for tracking the performance of hydraulic fracturing treatments within reservoirs. However, the low-magnitude earthquakes generated during fracturing and a large number of continuous raw records make the identification and detection of microseismic events a challenge. We treat the event detection as binary classification and utilize the machine learning techniques to build a model for identifying the events from noise. We also take advantage of the travel time and waveform reciprocities and develop the methods for source localization and focal mechanism inversions using a complex 3D earth model. Our methods can be used for processing the microseismic data in real-time.