I don' t understand why the same script with exactly the same amount of data (70Mb) is so slow in a more powerful server, anyone has an idea about what is the root of this problem?
I'm not familiar with that specific package but one explanation could be that the function is not parallelized so you effectively run the script in just one core regardless of how many you have available on the system and usually laptop cores are faster than server cores (individually). If this is the case you could take advantage of the extra cores in the server by manually parallelizing your code (if possible).
Can you provide more information about both systems? you said your laptop has 4 cores but they usually support 2 threads per core so you would get 8 "virtual" cores, matching the core count in the server so if the CPU frequency is higher in your laptop that would explain the faster performance.