Thread: Determine which stores to close

Hi,


Am not much of a stats guy so looking for help from you experts.


We have total sales data for around 100 stores for 12 months. The stores need to close for remodeling for 2 months; all the stores cannot be closed in the same month. There are few months were all the stores do good. In fact mostly the sales trend is uniform.
Any suggestion on best way to determine which stores should be closed when.


Currently, we have a very basic approach: determine the lowest 15% sales for a 2 month period (say Jan, Feb) and eliminate them, and then move to the next 2 month period...and continue.


Any advice? Any software , formula that will help?

Appreciate your help!
Thanks,
Bhavik

Hey barca.

You will need to first create a cumulative distribution for sales and identify the stores associated with each point.

Once you have the distribution, you get the point x where P(X < x) = 0.15 and all stores in that region will be in the bottom 50% range.

To do this get your sales data and build a histogram of it and use a statistical package to find what sales value the bottom 15% (or other percentage) is in.

Thanks a lot Chiro!
I obviously have to learn more about CDF and histogram to get a better understanding but I have a starting point now.

Pardon me, if my questions sound too trivial. Is there any software you suggest? I have used Matlab in the past so am fairly familiar with it.

I would suggest R and it is free as well:

The R Project for Statistical Computing

Thanks Chiro!