Setting Up Covariates for ARIMA Forecasting

[OneandOnly13]

Hi - I'm trying to create an ARIMA predictive model in Alteryx. I know that my data has seasonality to months, weeks, and day of the week. I'm wanting to setup my data to handle these as covariates. How would I structure my data so Alteryx can then use them in the TS Model Factory tool correctly? Below is a sample dataset and I'm wondering what additional columns are needed as covariates? I've also thought about forecasting these seasonalities as their own models and then blend them. What are the pros/cons to either method and if using the blended method, what would an example workflow look like?

Thanks in advance!

Date	Sales Channel	Type	Revenue
1/1/19	Web	New	$150
1/2/19	Web	Upgrade	$100
1/3/19	Web	Disconnect	($75)
1/4/19	Call Center	New	$125
1/5/19	Call Center	Disconnect	($50)
1/6/19	Store	New	$100

[jamielaird · Accepted solution]

Hi @OneandOnly13 

There is an example workflow for the TS Model Factory Tool in the Alteryx Gallery which may be useful to you:

https://gallery.alteryx.com/#!app/TS-Factory-Sample/5772b0ebaa690a1348cc6bcb

[OneandOnly13]

Thanks @jamielaird ! I can work with this example.