Online Retail Data exploratory analysis with Python

08 Apr 2025

CONCLUSIONS:

Addressing data management issues identified during the data cleansing step would be highly beneficial. If the dataset is to be maintained in Excel, consider adding macros to detect redundant observations and identify missing values in crucial fields like descriptions or customer IDs. Macros can also correct negative quantities, negative or zero unit prices, or unspecified countries.

Next, consider repeating the same analysis performed in this project for each of the 38 countries to obtain more detailed information about sales trends, customer behavior, and popular products. This will help tailor strategic business decisions and enhance the store’s overall performance.

Revising some of the descriptions to accurately capture the nature of the products in question is also recommended. Clearer descriptions would provide better insights into items like fees, postages, manuals, and adjustments for bad debt.

Further research should be conducted to identify the best products for countries with the smallest sales, such as Malta, the United Arab Emirates, or Saudi Arabia. Additionally, investigate why Sunday is the least busy day in terms of sales.

When forecasting future sales in the United Kingdom, be mindful of quantity outliers, as they appear to be the most sellable items.

Here is the Coursera Perform exploratory data analysis on retail data with Python completed project notebook.