Trend of Total Sales

[AdityaKir]

Hello,

My question is "In order to stock the inventory, the store wants to analyze the sales data and predict the trend of total sales for each dress for an extended period of three more alternative days."

For this the supporting data that I have got is as follows :

Dress_ID	29/8/2013	31/8/2013	2/9/2013	4/9/2013	6/9/2013	8/9/2013
1006032852	2114	2274	2491	2660	2727	2887
1212192089	151	275	570	750	813	1066
1190380701	6	7	7	7	8	8
966005983	1005	1128	1326	1455	1507	1621
876339541	996	1175	1304	1396	1432	1559

Not sure exactly which approach to use here or how to go about this problem.

Regards,

Aditya

[Kurt_Bremser]

The first thing to do is, of course, to transpose your dataset (once it has been imported into SAS; it can't be a SAS dataset at the moment as a date is not a a valid SAS column name) so you end up with three columns: ID, date, sales.

Then you need to calculate the difference from previous to current observation, so you get the daily sales; use the lag() function for this.

Once that is done, you can calculate average sales for certain timespans, or see how sales spread over a week, etc, etc.

[Reeza]

This is a time series analysis so review the Procs under SAS ETS for a relevant method. 

[gamotte]

Hi,

Since you have only a few observations at your disposal, you can proceed as follows

1. Transpose your dataset

2. Sum the columns to obtain total sales

3 . use a simple linear regression on a time trend with constant

4. Calculate the predicted values

[gamotte]

I missed the "for each dress" part of your question.
This kind of data is dealt with panel data methods.
Or, if you want to keep it simple, you can run a regression for each dress.

[Ksharp]

Yeah. Post it at Forecast Forum.
There are many PROC you could use like :  PROC ARIMA , PROC ESM .......