Programming with Python

Summary

The overall goal of the project is to utilize Python to process data, analyze data, visualize data and generate a business report about your analysis.

It is an individual project. You may discuss the logic with your peers, but you should complete the code and the report by yourself. Any identical submission will not be accepted.

Submission

You should submit 2 files:

.    A .ipynb file that includes your code and your report

.    A pdf version of your .ipynb file

Grading Criteria

1.   You should include necessary data processing steps (check data statistics, missing values, joins, column calculations)

2.   You should include your own analyses of the data (use calculations, use group by, aggregations, and visualizations)

3.   You should include narratives about your code and your analyses

4.   You should have clear conclusions based on your analyses

5.   You should provide recommendations for the business

6.   You should write high-quality code (including logic, error handling, and commenting)

Problem Description

Typically, websites have a clear path to conversion: for instance, you land on the home page, then you search, select a product, and buy it. At each of these steps, some users will drop off and leave the site. The sequence of pages that lead to conversion is called 'funnel'.

Funnel analysis allows understanding where/when our users abandon the website. It gives     crucial insights into user behavior and on ways to improve the user experience. Also, it often allows to discover bugs.

You are looking at data from an e-commerce website. The site is very simple and has just 4 pages:

.    The first page is the home page. When you come to the site for the first time, you can only land on the home page as a first page.

.    From the home page, the user can perform a search and land on the search page.

.    From the search page, if the user clicks on a product, she will get to the payment page, where she is asked to provide payment information in order to buy that product.

.    If she does decide to buy, she ends up on the confirmation page

The company CEO isn't very happy with the volume of sales and, especially, of sales coming from new users. Therefore, she asked you to investigate whether there is something wrong in the conversion funnel or, in general, if you could suggest how conversion rate can be improved.

Specifically, she is interested in:

.    A full picture of funnel conversion rate for both desktop and mobile (conversion probability for each page by device type)

.    Some insights on what the product team should focus on in order to improve conversion rate as well as anything you might discover that could help improve conversion rate. (extra credit may be awarded for high-quality insights and visualizations)