sales-predictions

A portfolio project demonstrating sales predictions for items sold at various retail outlets.

Business problem:

The goal is to help the retailer understand the properties of items and outlets that play roles in increasing sales.

Author: Olen Sluder

Data

The dataset consists of 8,523 rows of item and outlet data.

Column Name	Type	Description
Item_Identifier	String	Unique product ID
Item_Weight	Float	Weight of product
Item_Fat_Content	String	Whether the product is low fat or regular
Item_Visibility	Float	The percentage of total display area of all products in a store allocated to the particular product
Item_Type	String	The category to which the product belongs
Item_MRP	Float	Maximum Retail Price (list price) of the product
Outlet_Identifier	String	Unique store ID
Outlet_Establishment_Year	Integer	The year in which store was established
Outlet_Size	String	The size of the store in terms of ground area covered
Outlet_Location_Type	String	The type of area in which the store is located
Outlet_Type	String	Whether the outlet is a grocery store or some sort of supermarket
Item_Outlet_Sales	Float	Sales of the product in the particular store

Methods

• Missing numeric data is imputed using the mean value.
• Missing categorical data is imputed using the most frequent value.
• Grocery store type outlets with missing size are set to 'Small'.

Results

Does the item price and the outlet type impact sales?

• In addition to varying between outlet types, the correlation between item price and item outlet sales is not linear: there are items with high prices that do not sell well and items with low prices that sell well.
• The retailer may want to adjust the prices of certain items to increase their sales or promote higher-priced items with a strong sales record. They could also investigate why certain items with high prices are not selling well.

Do different types of outlets carry different types of items?

• The types of items sold varies across outlet types. In 'Supermarket Type1' outlets, the top-selling item types are 'Fruits and Vegetables', 'Snack Foods', and 'Household'. Other supermarket outlet types follow a similar pattern, but with lower total sales. In 'Grocery Store' outlets, the top-selling item types are 'Snack Foods', 'Fruits and Vegetables', and 'Household'.
• The retailer may want to promote the top-selling items in each outlet type to increase sales.

Model

A decision tree was selected for the machine learning model. It shows the ability to predict sales on a testing dataset to within an average of 1,057.43.

For further information

Please contact olen.sluder@gmail.com