Marketing Analytics Project
Marketing Campaign Analysis
Exploring how customer demographics, spending behaviour, and purchase patterns influence campaign response using Python, Pandas, and data visualisation.
Project Overview
Marketing campaigns often generate large amounts of customer data, but understanding which customers are most likely to respond can be challenging.
In this project, I analysed a public marketing campaign dataset to explore how demographics, household composition, spending behaviour, and recency relate to campaign response. The goal was not to build a predictive model, but to uncover useful patterns that could support better audience targeting and campaign planning.
Questions I Wanted to Explore
- Which customer groups had the highest campaign response rate?
- Does customer spending behaviour influence engagement?
- Do household characteristics affect response?
- Are recent customers more likely to engage with campaigns?
Dataset
This analysis uses a publicly available marketing campaign dataset containing customer demographics, purchase history, and campaign response information.
The dataset includes variables such as:
- Income and education level
- Marital status
- Number of children and teenagers in the household
- Customer recency
- Product spending behaviour
- Campaign response outcome
Approach
- Performed initial data exploration and quality checks.
- Handled missing values and removed inconsistencies.
- Created derived variables for customer segmentation.
- Calculated response rates across customer groups.
- Built visualisations to communicate findings.
- Translated observations into practical marketing insights.
Key Insights
High-spending customers responded more frequently
Response rates increased across spending segments, with the highest-spending customers showing the strongest engagement with marketing campaigns.
Recent customers were more engaged
Customers who had purchased more recently generally showed stronger response rates than customers with older purchase histories.
Household composition mattered
Customers without children or teenagers in the household tended to have higher response rates compared with larger family households.
What I Learned
This project was an opportunity to practise the complete analysis workflow: data cleaning, feature engineering, segmentation, exploratory analysis, and data storytelling.
More importantly, it reinforced that useful marketing insights often come from asking simple business questions and exploring the data systematically.
Explore the Full Analysis
The complete notebook, code, and documentation are available on GitHub.
← Back to Projects