Looping through each column and creating forecast models based on column name

[rravir]

Hello!

I am working on a dataset that has multiple columns. I want to create a workflow where I loop through each column, check the name of the column and create a forecast models based on the name. For example 

Period	Sales	Cost	Factora	Factorb
1	10	23	0.6	1.8
2	15	24	0.8	2.1
3	20	23	0.2	1.6
4	12	22	1.3	1.5
5	15	29	1.1	1.1
6	16	18	0.8	1.4
7	20	16	0.8	1.2
8	22	23	1.2	2.4

If the column name is Sales I want to create a univariate forecast model (ARIMA or ETS) and if the column is Cost I want to create a multivariate forecasting model (ARIMA) with Factora and Factorb as external variables. Is something like this possible to do in Alteryx?

I have been trying to use batch macros but with little success. Any help is greatly appreciated.  

[IraWatt]

Hey @rravir,

Have you checked out the TS Model Factory Tool & TS Forecast Factory Tool? I haven't personally used it but it allows you to generate multiple models based on groups of data which sounds like it would answer you problem. From what I know of it I think you would need to stack your target columns and add a group field for it to work but that should be alright with a couple of selects and then union tool.

If you have an issues setting up the factory tool make sure to ask :)

HTH,

Ira 

[leozhang2work]

As aforementioned, try out TS factory, it is what you are looking for.

Do make sure change the format of the data, so you can run it as group by.If you want to do different model, probably use a filter tool to split them.