Python-based exploratory data analysis to identify relationships between customer behavior, advertising spend, purchase activity, and annual revenue.
This project analyzes customer behavior data for NovaRetail, a retail company interested in understanding which variables are most strongly associated with annual revenue and customer activity.
The analysis focuses on identifying meaningful relationships between monthly purchases, monthly visits, targeted advertising spend, satisfaction, income level, customer age, and annual revenue. The objective was to use exploratory data analysis and statistical correlation techniques to support better business decisions around customer value, marketing investment, and future predictive modeling.
Retail companies generate large amounts of customer data, but not all variables contribute equally to business performance. Understanding which customer behaviors are most closely related to revenue can help teams prioritize marketing campaigns, improve customer segmentation, and identify high-value customer profiles.
For NovaRetail, this analysis helps clarify which variables are more relevant for explaining annual revenue and which variables may require deeper segmentation or advanced modeling before being used for decision-making.
Which customer behavior variables are most strongly associated with annual revenue?
Are monthly purchases and monthly visits related to customer value?
Is targeted advertising spend associated with higher customer engagement?
Do demographic or satisfaction variables show a strong relationship with revenue?
Which variables should be prioritized for future segmentation or predictive modeling?
The analysis found that monthly purchases are strongly associated with annual revenue. Customers with more monthly purchases tend to generate higher annual revenue, making purchase frequency one of the most relevant variables for identifying high-value customers.
Targeted advertising spend showed a moderate positive relationship with monthly visits, suggesting that marketing investment may be connected to higher customer engagement. However, this relationship should not be interpreted as direct causality without further validation.
Monthly visits also showed a positive relationship with both purchases and annual revenue. This indicates that visits can be useful as an early engagement signal, although they do not fully explain customer value on their own.
Variables such as age, income level, and satisfaction did not show strong linear relationships with annual revenue in this dataset. This does not mean they are irrelevant, but they may require segmentation, non-linear analysis, or more advanced modeling to reveal their business impact.
ConnectaTel could use the resulting customer segments to better align its plan offerings with actual customer behavior. High-usage customers may be strong candidates for premium or unlimited packages, while low-usage customers may respond better to basic plans, flexible pricing, or personalized promotions.
The company should also continue improving its data quality processes, especially around customer location, registration dates, and missing usage fields. Clean and reliable data is essential for building more advanced analytics models in the future.
As a next step, this analysis could be extended into a churn prediction model or a customer value segmentation dashboard to help business teams monitor customer behavior and identify commercial opportunities more effectively.
Data cleaning and validation with Python
Exploratory data analysis
Missing value treatment
Outlier analysis
Customer segmentation
Data visualization
Business-oriented interpretation
Communication of technical findings for stakeholders