Location-based services play an important role in Internet of Things (IoT) applications. However, a trade-off has to be made between the location estimation error and the battery lifetime of an IoT device. As IoT devices communicate over Low Power Wide Area Networks (LPWAN), signal strength localization methods can use the existing communication link to estimate their location. In this paper, we present a comparison of three proximity methods, one fingerprinting method and three ranging methods using Sigfox communication messages. To evaluate these methods, we use a ground truth Sigfox dataset which we collected in a large urban environment, as well as new evaluation data that was collected in the same urban area. With a mean estimation error of 586 m, our fingerprinting method achieves the best result compared to other signal strength localization methods.