Users from a streaming company can watch streams from both its app and website. We are given a dataset containing the following user information:
Time on App: average time spent on application, in hours/week;Time on Website: average time spent on the website, in hours/week;Length of Membership: how long a client is subscribed to service, in years;Yearly Amount Spent: average yearly expenditure, in dollars.
Understanding that the company wants to take action on maximizing profits, some questions arise from the dataset:
- Is the length of membership important?
- Should the company focus its investments on the app or the website?
- What other insights should be taken into consideration?
The dataset contains 500 non-null observations of the 4 user features.
The complete exploratory data analysis (EDA) notebook can be found here.
I also used the same dataset to explore Spark functionalities from reading and modifying the dataset to proposing a linear regression model. You can find this reference notebook here.
Solving the question:
Learning a new technology:
By analyzing the given data, one can arrive at the following conclusions, followed by some suggestions that might increase revenue:
Length of membershipis an important feature when determining theYearly Amount Spent. It implies that if the company had to focus on either creating new customers or retaining customers, the latter would be advisable because long-time users spend more on the service. On the same note, it might be interesting to display more advertisements for the same population strata.- The company should focus its investments on the app, seem that
Time on Apphas a much stronger correlation to our target.