Why are the values in the (python tutorial) response decomposition waterfall in the order of 1e20?

[leonl0000]

The response decomposition in cell 12 of the python tutorial notebook has values in the order of 1e20. The "trend", for example, is 455914982384.6B.

From my understanding, this is the amount a channel contributes to the dependent variable. But the dependent variable is revenue, and the dataset dt_simulated_weekly has revenue values in the single-digit millions. Checking robyn.model_outputs, I see that these values correspond to robyn.model_outputs.all_x_decomp_agg.xDecompAgg, but no aggregation of any values in the dataset could possibly get close to 1e20.

Am I misunderstanding something?

[alxlyj]

thanks for flagging this issue. we are looking into this and will provide an update once it's addressed. in the meantime, please use the R version as we investigate!