Response variable: Revenue
Scenario
Online shopper's purchase intention is a difficult problem to predict, given the very large number of factors that can potentially affect an individual's decision to buy or not buy an item. It gets worse because most online stores can only obtain analytics information to facilitate such predictions, but not full-fledged individual-level information. Nevertheless, determining primary factors related to the site visits which lead to an efficacious transaction occurring enables retailers to maximize site layout and marketing campaigns, in an attempt to
generate increased levels of sales.
This data set contains entries of a high volume of website visits, and the corresponding analytics information. The task is to forecast the value for the Revenue variable, with the models that can be applied to classification problems. Determine which features have the highest indication of a user's buying intention.
Attributes' Description
Administrative: Number of administrative pages visited by the user
Administrative_Duration: Length of time spent on administrative pages, measured in seconds
Informational: Number of informational pages visited by the user
Informational_Duration: Length of time spent on informational pages, measured in seconds
ProductRelated: Number of product related pages visited by the
user
ProductRelated_Duration: Length of time spent on product related pages, measured in seconds
BounceRates: Proportion of users who leave the site after only
interacting with a single page
ExitRates: Proportion of views to a particular page that were
the last in a user’s session
PageValues: Average number of pages visited by a user prior to
a transaction taking place
SpecialDay: Closeness of the visit to a notable day of the year
(e.g., Christmas Day, Mother’s Day, Valentine’s
Day)
Month: Month in which the site visit took place
OperatingSystems: Operating system of the user
Browser: Browser of the user
Region: Geographical region of the user
TrafficType: Traffic type of the user
VistorType: Whether the user is new or returning
Weekend: Whether the site visit took place on a weekend
Revenue: Whether the site visit resulted in a transaction
taking place