We have a set of products in which we are trying to determine which products we should continue to sell, and which products to remove from our inventory. The file contains BOTH historical sales data AND active inventory, which can be discerned with the column titled "File Type".

We suspect that data science applied to the set--such as a decision tree analysis or logistic regression, or some other machine learning model---can help us generate a value (i.e., probability score) for each product, that can be used as the main determinant evaluating the inventory. Each row in the file represents one product.

It is important to note that we have MANY products in our inventory, and very few of them tend to sell (only about 10% sell each year) and many of the products only have a single sale in the course of a year.

The file contains historical sales data (identified with the column titled File_Type) along with current active inventory that is in need of evaluation (i.e., File Type = "Active"). The historical data shows sales for the past 6 months. The binary target (1 = sale, 0 = no sale in past six months) is likely the primary target that should drive the analysis.

The other columns contain numeric and categorical attributes that we deem relevant to sales.

Note that some of the historical sales SKUs are ALSO included in the active inventory.

A few comments about the attributes included, as we realize we may have some attributes that are unnecessary or may need to be explained.

SKU_number: This is the unique identifier for each product. Order: Just a sequential counter. Can be ignored. SoldFlag: 1 = sold in past 6 mos. 0 = Not sold MarketingType = Two categories of how we market the product. This should probably be ignored, or better yet, each type should be considered independently. New_Release_Flag = Any product that has had a future release (i.e., Release Number > 1)

Your data will be in front of the world's largest data science community. What questions do you want to see answered?

(1) What is the best model to use that will provide us with a probability estimate of a sale for each SKU? We are mainly interested in a relative unit that we can continuously update based on these attributes (and others that we add, as we are able).

(2) Is it possible to provide a scored file (i.e., a probability score for each SKU in the file), and to provide an evaluation of the accuracy of the selected model?

(3) What are the next steps we should take?

If it is possible for me to send you the file, please let me know.

Thanks very much for any suggestions you may provide.

  • $\begingroup$ The question of which products to retire also depends on product substitution possibilities. Is there a possibility of substitution, or are all of the products truly independent? $\endgroup$ – Paul Dec 8 '16 at 15:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.