Subscribe to DSC Newsletter

Hi,

 

I have more than 100 stores; and there is a lot of store level and MSA level macro economic information. I need to forecast sales for each store for the next 1 year. And I just have SAS 9 (so no Proc Panel).

 

Been reading on Proc Mixed and Proc Tscsreg. Can anyone please guide me on which technique/procedure will be more appropriate and why? Any case study I can look up?

 

Also, which is a better approach? Forcast sales at the weekly level and derive the monthly sales? Or, forecast sales at the monthly level and derive the weekly sales?

 

Thanks,

Datalligence

Tags: Tscsreg, data, forecasting, mixed, panel, proc, series, time

Views: 290

Reply to This

Replies to This Discussion

Thanks Dirk. My understanding is that errors add up when we start with weekly level forecasts and derive monthly sales from them. I need to check more on this.

Based on my initial/little research:-) Proc Mixed seems to be the better option.
You should be using a Transfer Function model built for each store. With weekly data you can bring in 51 dummies for the seasonal effects. SAS can't do this for you automatically nor does it automatically adjust for outliers(trends, level shifts, outliers), etc.

Go here to read a case study using Transfer Function modeling(automated). http://www.autobox.com/pdfs/pos.pdf from Journal of Business Forecasting paper on "Forecasting and Planning with POS Data: A Case Study" (Winter 2008-2009).

You should use the weekly data if you have enough data to support it. The question whenever you forecast is "how fast can you react to the data?" If you have daily, but can only react at a weekly level then build the models at the daily and sum them to the weekly.

RSS

On Data Science Central

© 2019   AnalyticBridge.com is a subsidiary and dedicated channel of Data Science Central LLC   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service