Currently the code is both parsed and executed locally. For big data sets this will take a long time. The structure needs to be changed such that code is either executed on only the client side after it is deployed or is not executed in its entirety.