This thesis presents the application of signal processing techniques to rail transit data for the purposes of rail corrugation maintenance. There are three primary forms of data in this work to which signal processing techniques were applied: (1) the analysis of rail corrugation as a spatial signal, (2) the analysis of wayside noise and noise spectra, and (3) balancing time and frequency resolution for inspection of on-board vibration signals. Grinding maintenance is common for addressing deteriorating rail surface condition, as surface defects like corrugation are removed. However, measuring the rail surface condition and then making the actionable decision to grind can be complex. This work develops a summary of corrugation roughness RCIx(λ), an extension of RCI2018 which uses a representative block root-mean square (RMS) measurement to summarize corrugation across a segment. The act of measuring surface roughness is a time-consuming task for property maintainers, and the act of measuring rail that does not require maintenance is wasteful. The measurement and analysis of wayside noise can give insight into the state of the rail to help property maintainers determine whether rail in a particular location is believed to require direct measurement and intervention, thus minimizing wasted effort. Lveq(λ, v) is developed in this thesis as a noise weighting scheme meant to isolate corrugation-related noise from a wayside noise sample spectrum. The usage of vehicle-mounted accelerometers can provide massive amounts of up-to-date information on rail condition through regular revenue service. This scale of data collection can streamline measurement and maintenance decisions throughout a property, however, the information encoded in this data requires post-processing to identify the characteristics of interest. This work utilizes the wavelet transform as a time-frequency analysis tool to spatially locate a particular frequency attribute in an acceleration sample. The results presented herein showcase the effectiveness of the signal processing techniques applied to these data streams. The growth of corrugation and its response to grinding were demonstrated in both corrugation and wayside noise datasets, while the wavelet transform has demonstrated a sensitivity to characteristics of interest in on-board vibration data.