Bluetooth devices are a rich source of travel time data for transportation engineers. Like any other source, however, Bluetooth devices can generate outliers that may bias travel times and corridor speeds. In this study, the proposed, innovative statistical methodology is capable of real-time deployment when travel time and vehicle speed outliers collected from Bluetooth data collection systems are detected. The proposed statistical methodology identifies outliers by using a data point's standard residual in a robust Greenshields model. The efficiency and the feasibility of this statistical methodology are evaluated for both Interstate highways and urban arterial corridors. These roadways are selected because they display a wide range of traffic patterns and outlier generators. The effectiveness of one methodology, especially on a wide range of traffic patterns, would improve the potential for widespread implementation. Four subsets of the collected data are provided within this study to highlight the efficiency and capability of the methodology. The results of using the Shapiro–Wilk statistics show that in both Interstate highway and urban arterial deployment the proposed method is effective at identifying Bluetooth outliers and is capable of working in a real-time environment. A second comparison between the two road types shows that the methodology performs better for Interstate highways than urban arterials, for which the added performance is based on the higher volume of hits and less fluctuation.