Unfortunately I can't even replicate it in Expert mode without R...
We are trying to build a tool for energy forecasting.
The complexity comes from the fact that we need to create a separate forecast for each of around 1000 customers.
So the exponential smoothing algorithms are out of question because they only process a single time series (and because I have multiple customers, the dates get repeated).
Because of this the automated analytics tool gives me an error because it finds duplicate dates (although they are not duplicates, just same date for a different client).
This is why I decided to go the R route.
I guess it would be useful to have some kind of segmentation option for the algorithms (for instance, the exponential smoothing can have the option to segment by customer/region/whatever so that a different model is created for each customer/region/etc).
In the example I've sent you got the error just by getting a date from the Excel source, converting it to the 2015-05-25 format for R and then converting it to string - also for R as it seems the custom components cannot accept a date field as an input.
Thank you for all your help!
Dan