how to forecast best weekly food menu

[HitmonTran]

Hi , New to SAS.

I need to create a calendar that has set food menu from different vendors for the entire month, the following month, and so on. The tricky part is finding out which food vendor works well with other vendors (eg. Bob's Burger + Mr. French Fries would work better than Bob's Burger + McDonalds) and which days work better for that vendor than other days.

I was wondering if there's a package or script that can process my raw data and forcast them to see:

1. which vendor should be grouped with which other vendors for which week-date(s) for which location(s)
2. create Auto-calendar generation that maximizes sales
3. Number of orders per vendor per location per week on average (basically removing bias)
4. Determine optimal rotation "cadence" per vendor (bi-weekly, tri-weekly, etc.)

I currently have 6 months worth of data that consists:

1. Order date (including day and time)
2. Food Vendor
3. Food item
4. Customer name
5. Location

So far I am able to get simple frequency counts on days, food item, food vendor, but i need more insight than overall frequency count. Maybe find all of the correlations between the different vendors. That way you can pair them together for the schedule. 

Let me know if you need more info, I am stuck here. Thanks!

[Reeza]

Google Market Basket Analysis - it's an approach to determine which items people tend to buy together. 

---

@HitmonTran wrote:

Hi , New to SAS.

I need to create a calendar that has set food menu from different vendors for the entire month, the following month, and so on. The tricky part is finding out which food vendor works well with other vendors (eg. Bob's Burger + Mr. French Fries would work better than Bob's Burger + McDonalds) and which days work better for that vendor than other days.

I was wondering if there's a package or script that can process my raw data and forcast them to see:

1. which vendor should be grouped with which other vendors for which week-date(s) for which location(s)
2. create Auto-calendar generation that maximizes sales
3. Number of orders per vendor per location per week on average (basically removing bias)
4. Determine optimal rotation "cadence" per vendor (bi-weekly, tri-weekly, etc.)

I currently have 6 months worth of data that consists:

1. Order date (including day and time)
2. Food Vendor
3. Food item
4. Customer name
5. Location

So far I am able to get simple frequency counts on days, food item, food vendor, but i need more insight than overall frequency count. Maybe find all of the correlations between the different vendors. That way you can pair them together for the schedule. 

Let me know if you need more info, I am stuck here. Thanks!

---