The goal of this study was to take the first step towards a diagnosis tool for multi-joint patients by investigating whether it is possible to distinguish subjects respective their disease group solely based on discrete parameters describing their gait pattern. A dataset including patients with unilateral knee or hip osteoarthritis, cervical or lumbar spinal stenosis, and healthy controls was compiled. All participants performed a gait analysis comprising approximately 30 steps at their self-selected preferred walking speed. Spatiotemporal parameters and ankle, knee and hip kinematic trajectories were captured using the inertial sensor system RehaGait® (RehaGait Hasomed GmbH, Magbedurg, Germany). Data of 254 participants were extracted from the laboratory database projects. After checking for duplicates and removing data that did not pass the quality check, 206 participants remained in the final dataset, including 30 patients with unilateral knee ostoarthritis, 20 with unilateral hip osteoarthritis, 43 with lumbar spinal stenosis, 17 with cervical spinal stenosis, and 96 healthy controls. Discrete parameters were extracted for each step using an inhouse algorithm written in Matlab. Mean values for all parameters were calculated for each participant and side and used for further analysis. Sociodemographic data, spatiotemporal parameters, and calculated angles were collected in one dataset, which was then checked for missing values. Missing values of the healthy participants were filled with the mean angle of this group. Missing values in the patient group were filled using the nearest-neighbors method.