The measurement campaign was performed in Cologne (Germany) using a RUSK MIMO channel sounder operating at 2.53 GHz. The fixed transmitter has been mounted on a roof top of an approx. 30m high building whereas the mobile receiver has been set up in a car. The application of high resolution parameter estimation algorithms requires adequate antenna arrays. Therefore the measurements have been done using a stacked uniform circular array at the base station side (Tx) as well as at the mobile station side (Rx). Both arrays are linear dual polarized. In order to obtain an antenna independent characterization of the radio channel the high resolution parameter estimation framework RIMAX was used. RIMAX models the channel as the superposition of a deterministic (specular) and a stochastic (dense multipath) part. The deterministic part resembles a multitude of plane waves that are resolvable by the estimator. Each plane wave is hereby described using its time-delay, angle-of-arrival (in azimuth and elevation), angle-of-departure (in azimuth and elevation) and the complex polarimetric path-weight. In order to render the data self contained, we have also added suitable calibration data of both antenna arrays. Only through the angle-, polarization- and port-dependent description of the antenna arrays one can achieve correct deembedding of the measurement equipment. In addition to the measurements and calibration data, we also supply four different sets of high resolution parameter estimates that are derived form the data based on RIMAX where different strategies for model-order selection and over-equalization were employed. Since these different settings impose different objective functions and parametric models onto the data, the resulting parameter estimations have signficantly different statistics, which sparks the need for further analysis that is facilitated by this dataset.