The ability of cells to migrate towards chemoattractants is essential for many (patho- )physiological processes such as wound healing or tumor metastasis. Chemotaxis is transduced by complex molecular signaling cascades coupling to the intrinsic cell migration machinery. The overall effect is typically assessed with time- lapse video microscopy of single migrating cells. Interfering with the signaling cascade allows to relate microscopic processes with the movement of the whole cell. However, typical quantities such as mean velocities or the mean chemotactic index that can be calculated from the experimental cell paths provide only a coarse interpretation of the migratory behavior. Thus, we analyze all experimental data within the concept of stochastic processes to gain further insight into the biological activities. The cell is regarded as an object driven by internally correlated stochastic forces and external fields generated by chemoattractants. The analysis of migration of MDCK-F cells without external stimuli revealed a variety of anomalous properties being due to strong temporal correlations that could be described with so-called fractional equations. Anomalous dynamics is preserved when MDCK-F cells are exposed to a stable gradient of chemokine fibroblast growth factor 2. We extract further quantities from the experimental data - especially probability density functions (pdf) of direction, position and orientation with kernel-based methods – to classify in more detail the chemotactic response and the influence of the TRPC1 channels therein. Bayesian data analysis is able to select the best model pdf for a given set of experimental data. The numerical simulation of different stochastic models allows assessing the quality of the different parameters. We finally discuss possible links between the observed strong temporal correlations and appropriate responses to chemoattractants.