Through the last year’s SURF project, we built a multivariate dataset—i.e., XJTLUIndoorLoc—for indoor localization and trajectory estimation based on Wi-Fi received signal strength (RSS) and geomagnetic field, which covers the 4th and the 5th floor of the IBSS building and includes measurement data at 969 reference points. In this project, we are to carry out a systematic analysis of XJTLUIndoorLoc dataset to investigate the issues of the dependency of measurement data on mobile devices and the lack of device orientation information for geomagnetic field in deep neural network (DNN)-based indoor localization and trajectory estimation.