Hi,
That's exactly what planning functions do - manipulate the data before writing to the target real time cube.
You can create a sequence of manipulations, for example step 1 take the data from "side" cube and nultiply by 10, step 2 take the data from real time and divide by 10, step 3 add first dataset and second dataset, step 4 run over unified dataset and do something, step 5 save the data.
BR,